Factor analysis as a tool to explore the heterogeneity of type 2 diabetes: A feasibility study

Diabetes is currently classified based on pathogenetic mechanisms and Type 2 diabetes (T2DM) can be considered a residual heterogeneous category. Factor analysis (FA) identifies a limited number of calculated variables related to a larger number of measured parameters, capable of explaining most of their variance. Aim of the present study was to verify the feasibility of the application of FA for the development of pathogenetic models of individual cases of T2DM, using three available databases.

Firstly, a model of FA was applied to an existing dataset of non-diabetic patients, identifying three factors associated with fasting or post-prandial hyperglycemia. These factors were then calculated in: - patients enrolled in a retrospective observational study, assessing time to failure to diabetes treatment in three cohorts of patients (metformin or sulfonylurea monotherapy, or no pharmacological therapy); - in a retrospective cohort of patients failing to dual oral therapy and initiating treatment with DPP4 inhibitors; - in patients enrolled in a case-control study onincident cancer in T2DM subjects initiating insulin treatment.

Despite limitations, our results confirm the feasibility of approaching the characterization of T2DM through the identification of dimensional factors, providing additional and complementary information to that obtained with cluster analysis.