Socioeconomic modification of the diet quality-HbA1c association among U.S. adults: A survey-weighted interaction analysis

1.Traditional nutrition science often proceeds under the assumption of a universal metabolic return to healthy eating, yet social environments may fundamentally modify these biological associations. This investigation utilized survey-weighted data from the National Health and Nutrition Examination Survey (NHANES 2017-2023) representing a weighted population of 286 million adults aged 20 years and older to test for association heterogeneity in the dietglycemia relationship. Dietary exposure was operationalized as energy-adjusted nutrient density scores derived via the residual method to measure the healthfulness of intake independent of total caloric volume. The primary outcome was HPLC-measured glycated hemoglobin (HbA1c) modeled as a continuous variable. Multivariable interaction models evaluated the Income-to-Poverty Ratio (PIR) as a formal effect modifier, adjusting for age, sex, race/ethnicity, body mass index, and smoking status. Analysis demonstrated that while quality-weighted nutrient intake levels remained statistically uniform across income tiers (p=0.207), multivariable interaction models identified significant modification of the diet-HbA1c association by socioeconomic position (p=0.028 for interaction). In the low-income reference group, higher quality nutrient intake was associated with a significant protective decline in HbA1c ({beta}=-6.11x10-5 percentage points per kilocalorie, p=0.017). Conversely, this protective association was attenuated to non-significance for the middle-income stratum ({beta}=3.31x10-6, p=0.849). Sensitivity analysis restricted to participants without clinical diabetes showed the interaction was non-significant (p=0.859), identifying an observed boundary condition where differential associations are most evident in pathological states. These findings suggest association heterogeneity where identical dietary behaviors relate to divergent glycemic patterning across socioeconomic groups. Socioeconomic position appears to function as an effect modifier of the diet-HbA1c relationship rather than a mere confounder. Such patterns suggest that the metabolic correlates of healthy eating are socially patterned, potentially due to structural factors that constrain the associations expected from improved nutrition.