MODERATOR ANALYSES OF THE COMPARISON BETWEEN REAL-TIME AND INTERMITTENTLY SCANNED CONTINUOUS GLUCOSE MONITORING IN ADULTS WITH TYPE 1 DIABETES: A SUB-ANALYSIS OF THE ALERTT1 TRIAL

Abstract

Background and Aims: The increasing use of continuous glucose monitoring (CGM) devices in individuals with type 2 diabetes (T2D) opens the door to the use of patient generated data to track disease progression and/or therapy effectiveness. The key parameter quantifying the quality of glucose control is the disposition index (DI), the product between insulin sensitivity (S I) and residual beta-cell function (F). Different methodologies were proposed in the literature for DI calculation, like the Oral Minimal Model (OMM), but all require collecting plasma measurements during hospitalized setting. Here we propose a method to quantify DI in everyday life conditions, using CGM data, and validated it against OMM. Methods: The method was tested in silico using the 100 virtual subjects of the Padova T2D Simulator (Visentin et al., 2020). Two single-meal scenarios (75g carbohydrates) were performed, with subjects receiving either placebo or a hypothetic treatment, designed to improve both S I and F by 50%. Plasma glucose, insulin and C-peptide concentrations were measured for 6 hours after each meal and used for the estimation of the reference DI with the OMM (DI MM), while CGM sensor data were employed for the calculation of DI with the sensor-based method (DI SB). Results: DI SB well correlated with DI MM and both were able to significantly detect (p<<0.001) DI improvement in treatment vs placebo (Fig. 1). Conclusions: The DI SB can be used to assess DI in subjects with T2D wearing CGM. Future work will include testing the method on real data of T2D with different stage of disease progression. Background and Aims: ALERTT1 showed that switching from isCGM to rtCGM with alert functionality improved time in range (TIR; 70-180 mg/dL), HbA1c, time <54 mg/dL, and Hy-poglycemia Fear Survey worry score (HFS-worry) in adults with type 1 diabetes (T1D). It is not known whether certain subgroups benefit more from switching to rtCGM than others. Methods: This post-hoc analysis of ALERTT1 verified whether the 6-month difference in means of rtCGM versus isCGM (delta) for TIR, HbA1c, time <54 mg/dL, and HFS-worry depended on different patient characteristics, by including the interaction of 14 variables with the delta in a moderator analysis. Analyses were performed for each of these variables separately (univariable analysis); variables with p < 0.10 in the univariable analysis were combined into a single model (multivariable analysis). Results: Univariable analyses showed no statistically significant dependency of delta TIR on another variable; only dependency of delta HFS-worry on HbA1c was observed (p = 0.0059), indicating less worry with the use of rtCGM in people with low (£6.5%) or high (‡8%) baseline HbA1c. Given p < 0.10 for the dependency of delta TIR on insulin therapy (p = 0.0851; favoring multiple daily injections), baseline HbA1c (p = 0.0537), and baseline TIR (p = 0.0615), these variables were combined into a multivariable analysis. None of the interactions were statistically significant. Conclusions: Except for HFS-worry, no interactions between 14 variables and the 6-month intervention effect of rtCGM on TIR, HbA1c, or time <54 mg/dL were observed, supporting the conclusion of ALERTT1 that the benefit of switching from isCGM to rtCGM with alert functionality applies to a wide range of people with T1D.