A multidimensional analysis of CGM-derived glucose metrics in type 1 diabetes: differential associations of insulin degludec and insulin glargine U300

Abstract

Ultra-long-acting basal insulins, insulin degludec (IDeg) and insulin glargine U300 (IGlarU300), are widely used in type 1 diabetes mellitus (T1DM), but their differential effects on multidimensional continuous glucose monitoring (CGM) profiles remain unclear. We aimed to identify latent dimensions underlying CGM-derived metrics in T1DM and to compare these dimensions between IDeg and IGlarU300. This predefined secondary analysis used data from the multicenter, randomized, crossover KOBE-BBIS2 trial in C-peptide-negative individuals with T1DM. Participants received IDeg and IGlarU300 for 4 weeks each, and professional CGM was performed during the final week of each treatment period. Exploratory factor analysis was applied to conventional CGM metrics and two autocorrelation-based indices (AC_Mean and AC_Var). Associations between clinical characteristics and factor scores were also examined. Three latent components were identified: high-glucose/high-variability, autocorrelation, and low-glucose/high-variability. The low-glucose/high-variability factor score was higher during IDeg treatment than during IGlarU300 treatment (mean difference 0.40, 95% CI 0.01–0.78). In exploratory analyses, body mass index was inversely correlated with this factor during IDeg treatment (r = − 0.45, 95% CI − 0.73 to − 0.11). These findings require validation in larger independent cohorts.