Time in Range and Incident Diabetic Retinopathy: Clinical and Economic Evidence From Real-World Type 1 Diabetes Care.

Abstract

Continuous glucose monitoring (CGM) metrics are increasingly used to study the relationship between glycemic control and diabetic complications; however, the correlated glucose measurements and unequal follow-up times in real-world CGM data necessitate more advanced analytical approaches to yield unbiased estimates. This retrospective case-cohort study aimed to estimate longitudinal changes in time in range (TIR) on progression to diabetic retinopathy (DR) among patients with type 1 diabetes (T1D). This was a retrospective case-cohort study among patients with T1D, using CGM devices. We analyzed linked CGM and electronic health record data from 161 patients with T1D, with long-term follow-up for incident DR diagnoses. TIR was defined as time spent in sensor glucose between 70 and 180 mg/dL. Multilevel mixed-effects parametric survival models and Markov models were constructed to obtain effects of TIR (eg, hazard ratios [HRs]) and lifetime trajectories of developing retinopathy and blindness, respectively. A retrospective case cohort of 161 patients with T1D (mean duration 13.7 years) included 71 cases (baseline glycated hemoglobin A1c [HbA1c] of 8.2%) and 90 controls (baseline HbA1c of 7.3%). A 10% increase in TIR was associated with a modestly lower risk of retinopathy progression (HR 0.88; 95% CI, 0.78-0.98) and an estimated prevention of 39 cases of blindness per 1000 individuals over time (TIR 70% vs 40%). Economic simulation modeling suggested $2581 lower costs per person and a gain of 0.13 quality-adjusted life years (QALYs). These findings can guide real-world CGM studies and support diabetes simulation models to predict future treatment outcomes from CGM metrics.