CGM-derived disposition index estimates beta cell function in individuals with islet autoimmunity who do not require insulin treatment; changes in the disposition index precede changes identified by CGM metrics 1 year later

Abstract Aims/hypothesis Monitoring of beta cell function in individuals with islet autoimmunity who do not require insulin treatment (preclinical type 1 diabetes) is normally done via OGTTs, posing major challenges given the invasiveness of the test. Here, we validated a disposition index (DI) derived from continuous glucose monitoring (CGM) metrics (DI CGM ), which is a proxy of beta cell function, against the gold-standard OGTT DI (DI MM ), in a cohort of individuals with one or more islet autoantibodies who do not require insulin treatment. Methods Participants underwent a 3 h 10-point OGTT with measurement of glucose, insulin and C-peptide while wearing a blinded CGM. The sensor was started 24 h prior to the OGTT. DI CGM was computed using CGM data during the 3 h OGTT and two blood glucose calibrations, and compared with the DI from the oral minimal model (DI MM ), which uses plasma glucose, insulin and C-peptide. The association between the baseline DI CGM and duration of glucose levels above 7.8 mmol/l (140 mg/dl) (TA140) at 1 year was explored. Results Twenty-four participants (median age 16.8 years [IQR 13.5–27.7] ) were enrolled, of whom 18 completed both the OGTT (ten with pre-stage 1, seven with stage 1 and one with stage 3a type 1 diabetes) and CGM at baseline, and had repeat CGM 1 year later. The baseline DI CGM was highly correlated with the DI MM ( r =0.90, p <0.001). Lower baseline DI CGM was associated with higher 1 year TA140 ( r =−0.61, p =0.014) and higher 1 year mean sensor glucose ( r =−0.62, p =0.020). Conclusions/interpretation Beta cell function can be estimated using CGM during a glucose load. DI CGM correlates with a 1 year clinically relevant outcome (TA140) in individuals with islet autoimmunity who do not require insulin therapy.