1596-P: Personalized Glycemic Response to Carbohydrates and Associated Physiological Signatures in Multiomics

Introduction & Objective: Individuals with type 2 diabetes (T2D) exhibit heterogeneous metabolic dysfunction (e.g., muscle insulin resistance (IR), beta-cell dysfunction) and varying intervention needs. People also have different dynamic responses to daily activities (e.g., exercise and food), which are associated with their metabolic function. Understanding personalized glycemic control and specific metabolic dysfunctions is crucial for T2D subtyping and personal intervention.

Methods: Here, we incorporated real-life continuous glucose monitoring (CGM) to evaluate the personal responses to 7 different carbohydrate foods (each 50 grams of carbohydrate) and 3 different nutrients (fat, protein, and fiber) that tend to mitigate glycemic spikes in 50 participants (25 with prediabetes and 25 with normoglycemic levels). We built a computational workflow to extract 20 different metrics (e.g. peak height and time to peak) related to metabolic functions, subtype participants according to their relative response to different foods, and uncover physiological and pathological associations in multi-omics.

Results: We found that the average glycemic response depends on carbohydrate types and fiber contents. Worst metabolic control (e.g., IR) is also significantly associated with higher glycemic spikes. In addition, we discovered different CarbResponders, that different groups of people have different relative glycemic spikes to different foods. Some people spike more on rice and others on grapes. We found a higher proportion of rice spiking in Asians (p=0.007), better metabolic functions (e.g., IR) associated with lower spikes in starch food, and signature of diet habits in metabolomics and proteomics.

Conclusions: Food response profiles provide individualized food recommendations and indicate underlying metabolic functions for personalized interventions.