Amber Sabbe, Niels Bochanen, Annelies Van den Broucke, Katarina Mihaljčić, Axelle Bigaré, Kristof Van Dessel, Jason Bouziotis, Kristien Ledeganck, Eveline Dirinck, Nancy Bolsens, Jolien De Meulemeester, Pieter Gillard, Christophe De Block
CGM-derived glucometrics were associated with the 5-year incidence of microvascular complications in this real-world cohort, with ΔTIR potentially providing additional prognostic information beyond HbA1c. Further validation is needed before routine clinical implementation.
CONTEXT: Beyond HbA1c, continuous glucose monitoring (CGM) is now widely used in clinical practice, providing more detailed information about glycemic control. However, its ability to predict long-term microvascular outcomes in type 1 diabetes remains unclear.
OBJECTIVE: To evaluate the association between CGM-derived glucometrics and the 5-year incidence of microvascular complications, providing longitudinal real-world evidence in adults with type 1 diabetes.
DESIGN: Longitudinal real-world cohort study with five-year follow-up.
SETTING: Tertiary referral center.
PARTICIPANTS: 246 adults with type 1 diabetes (median age 40 [28-53] years, diabetes duration 24 [13-33] years).
INTERVENTION: No interventional treatment.
MAIN OUTCOME MEASURE: Composite microvascular endpoint including nephropathy (persistent albuminuria ≥20 µg/min, eGFR <60 mL/min/1.73 m2, and/or RAAS inhibitor use for albuminuria), severe retinopathy (proliferative retinopathy, maculopathy, and/or blindness), and peripheral neuropathy (monofilament score <6/6 and/or abnormal nerve conduction).
RESULTS: HbA1c decreased from 7.6% (59 mmol/mol) to 7.2% (55 mmol/mol), and time in range (TIR, 70-180 mg/dL) increased from 50 ± 14% to 61 ± 15%. All CGM-derived metrics (TIR, mean glucose, glucose management indicator) were associated with the composite endpoint. In this longitudinal real-world cohort, a ≥ 10% decrease in TIR was associated with higher complication rates (33.3%) compared with a stable TIR (7.9%) or ≥10% increase (15.0%) (p = 0.011). In multivariable analyses adjusted for age and baseline TIR, change in TIR (ΔTIR) remained independently associated with outcomes. BMI and systolic blood pressure were not significant. Models incorporating glucometrics and age showed acceptable discriminative performance (AUCs 0.757-0.787).
CONCLUSIONS: CGM-derived glucometrics were associated with the 5-year incidence of microvascular complications in this real-world cohort, with ΔTIR potentially providing additional prognostic information beyond HbA1c. Further validation is needed before routine clinical implementation.