A family-based study links greater intake of ultra-processed foods with early changes in insulin-related measures in nondiabetic children, underscoring the need for longer-term research.

Study: Association of Ultra-Processed Food Intake with Insulin Secretion and Sensitivity in Mexican American Children: Results from the SAFARI Study. Image Credit: Rimma Bondarenko / Shutterstock

A new study published in the journal Nutrients examined whether ultra-processed food intake was associated with insulin secretion and sensitivity.

Ultra-Processed Foods and Their Metabolic Implications

Ultra-processed foods (UPFs) are highly industrially processed products containing ingredients and additives not commonly found in home kitchens, such as hydrolyzed proteins, high-fructose corn syrup, and cosmetic additives. Over the past decade, UPF consumption has risen dramatically worldwide, most notably in high-income countries, but also in middle- and low-income regions, driven by urbanization and shifts in food industry policies.

Increasing UPF intake has been associated with heightened cardiometabolic risk through several mechanisms. UPFs are often high in carbohydrates, low in fiber, hyper-palatable, and rapidly digested. They may be associated with changes in gut microbiota, intestinal permeability, and chronic low-grade inflammation, all of which may drive metabolic stress and impair glucose regulation. Furthermore, frequent consumption of rapidly absorbable carbohydrates and the lower satiety provided by UPFs may negatively affect insulin action and the pancreatic response to food.

Most studies on UPF intake focus on late-stage outcomes like obesity and diabetes, overlooking early changes in insulin sensitivity and secretion that can be detected with oral glucose tolerance tests (OGTTs). Research on the early metabolic effects of UPFs in children, especially regarding insulin sensitivity and beta-cell function, remains limited. Addressing this gap may help clarify whether UPF consumption is associated with changes in children’s metabolic health before clinical disease emerges.

Researchers Examined Dietary Patterns and Insulin Sensitivity in SAFARI Youth

The current study is a secondary analysis of the San Antonio Family Assessment of Metabolic Risk Indicators in Youth (SAFARI) study. SAFARI included children and adolescents aged 6–17 whose adult family members participated in one of three genetic epidemiology studies: the San Antonio Family Heart Study (SAFHS), San Antonio Family Diabetes/Gallbladder Study (SAFDGS), or Veterans Administration Genetic Epidemiology Study (VAGES).

Most families were lower-income Mexican Americans from San Antonio, enrolled between September 2005 and May 2010. The study included 673 participants from 401 nuclear families, averaging two children per sibship. In the original SAFARI cohort, youth exhibited high rates of obesity, metabolic syndrome, and prediabetes.

The current study used available data from participants with dietary and insulin secretion or sensitivity measures, with sample sizes varying by outcome. Nineteen indices of insulin sensitivity and secretion were calculated, including 10 estimated from OGTT results. Dietary intake was assessed using the Block Kids’ Food Frequency Questionnaire (FFQ) in 508 children, analyzing responses to 78 food and beverage items. UPF intake was estimated from the frequency and quantity of 61 listed items.

Higher UPF Consumption Associated With Insulin Dysregulation

The study participants had an average age of 11.5 years, 49% were female, and the mean body mass index (BMI) was about 22.3 kg/m². More than 30% of the cohort were classified as having general or abdominal obesity. Prediabetes, defined as impaired fasting glucose, impaired glucose tolerance, or both, was present in 13% of participants, high blood pressure in 12%, dyslipidemia in 32%, and metabolic syndrome in 19%.

The family-based analysis accounted for relatedness among 3,664 kinship pairs, including third cousins, second cousins, first cousins, and siblings. There were notable differences between males and females in systolic blood pressure, physical activity, fasting glucose, and measures of insulin function, including HOMA_β (homeostasis model assessment of beta-cell function) and several OGTT-based insulin indices, such as Matsuda, insulinogenic, and disposition indices.

The analysis showed that several insulin-related traits had substantial genetic contributions, supporting the use of polygenic models to account for relatedness in the association analyses. Most continuous traits had heritability estimates above 0.43, while estimates were not reliable for serum C-peptide or ISI0,120.

UPF made up 53.3% of total energy intake among participants. After adjusting for confounders, a high UPF score was significantly associated with seven of 19 insulin-related traits at the conventional p \< 0.05 threshold. Higher UPF consumption correlated with lower insulin sensitivity, as measured by HOMA_s (homeostasis model assessment of insulin sensitivity) and QUICKI. It was also linked to increased fasting insulin and greater insulin resistance, as indicated by HOMA_IR.

With regard to OGTT-based indices, increased UPF intake was associated with lower Matsuda index values and higher values for both the insulinogenic index at 30 minutes and the disposition index at 30 minutes. Nominal associations, with p-values between 0.05 and 0.10, were also reported for HOMA_β, the area under the curve for insulin (AUC_Insulin), and the triglyceride-glucose index (TyG). Collectively, these results suggest that higher UPF scores were associated with several measures of insulin sensitivity, insulin resistance, and early insulin secretion in nondiabetic children. However, the study did not correct for multiple comparisons across 19 traits, meaning that its findings are exploratory and indicative rather than confirmatory. Using the Bonferroni threshold discussed by the authors (p \< 0.0026), only the 30-minute insulinogenic index remained statistically significant.

Adjusting for total calorie intake attenuated and made nonsignificant associations between UPF intake and insulin-based fasting indexes, including fasting insulin, HOMA measures, and QUICKI. Conversely, associations with fasting glucose-based indexes became stronger and significant. Among OGTT-related indexes, only the Matsuda index lost significance, while the insulinogenic and disposition indexes at 30 minutes remained significant. This pattern suggests that total energy intake may explain part of the observed associations, but it cannot establish calorie-independent effects or underlying mechanisms.

Findings in Nondiabetic Youth Need Confirmation

The current study indicates that higher UPF intake was associated with altered measures of insulin secretion and sensitivity, a finding observed even in young individuals after accounting for measured covariates and pedigree structure, although gene-environment interactions could not be excluded. These observations underscore the potential role of early dietary patterns in shaping long-term metabolic health and suggest that early interventions to reduce UPF consumption could inform prevention research and future randomized intervention trials.

However, this cross-sectional analysis cannot establish that UPF intake causes insulin dysregulation. Its reliance on a food-frequency questionnaire not designed to quantify UPF intake, possible misclassification of UPFs, smaller and varying sample sizes for OGTT outcomes, and dietary data collected more than 20 years ago also limit interpretation and generalizability. Future studies, particularly those involving larger, longitudinal populations and randomized trials, are warranted to confirm these associations and unravel the underlying biological mechanisms. The authors plan to recall the SAFARI study cohort to evaluate the persistence and possible progression of these metabolic effects over time.

Journal reference:

• Kulkarni, H. et al. (2026). Association of Ultra-Processed Food Intake with Insulin Secretion and Sensitivity in Mexican American Children: Results from the SAFARI Study. Nutrients. 18(14), 2361. DOI: 10.3390/nu18142361, https://www.mdpi.com/2072-6643/18/14/2361