Deep digital and immuno-phenotyping of people with type 1 diabetes according to their metabolic and psychological stress levels for precision prevention of diabetes-related complications.

Background: Type 1 diabetes (T1D) is a frequent autoimmune disorder resulting from destruction of the β-cells in pancreatic islets1. Besides glycemic instability, mental burden, diabetes distress, fatigue, depressive symptoms are key risk factors for complications but little is known about their immunological impact.
Objectives: The project is divided into 3 objectives: 1) Deep digital phenotyping and clustering of people with T1D according to their levels of psychological and metabolic stress. 2) Deep immuno-phenotyping of people with T1D according to their levels of psychological and metabolic stress and study of the associations between immune and digital phenotypes. 3) Identification of clinically relevant deep digital and immuno-phenotyping clusters and study of their associations with diabetes-related microvascular complications and intermediate markers of cardiovascular health.
Material & Methods: The project is based on the SFDT1 study where up to 15,000 participants with T1D are currently being recruited in France. 1) Deep digital phenotyping will be performed on both data from continuous glucose monitoring devices and from e-patient reported outcomes of diabetes distress and anxiety in N=1500 participants. 2) Deep immuno-phenotyping will be based on frozen plasma blood samples (antibodies [ICA, AA,GAD, IA2/ICA512], cortisol, catecholamines, us CRP, IL-1ꞵ, IL-6, IL-17, TNF𝛂, MCP-1, anti-microbiota IgG, monocytes) in N=1000 participants. Machine learning and clustering techniques will be used to derive patterns of digital and immuno-phenotyping that are clinically relevant. Logistic regression models will also be employed to study the associations with the main diabetes-related complications.