From Clinician-Based Risk Score to Diabetes Complications: Validating the Glycemia Risk Index with the Virtual Diabetes Control and Complications Trial Continuous Glucose Monitoring Data.

BACKGROUND: The glycemia risk index (GRI) is a validated continuous glucose monitoring (CGM)-derived metric of overall glycemic risk. Hemoglobin A1c (HbA1c) has been the gold-standard metric of glycemic control since the Diabetes Control and Complications Trial (DCCT). This study evaluates the relationship of GRI with long-term diabetes complications and compares its performance with HbA1c within the Virtual DCCT. METHODS: The Virtual DCCT reconstructed CGM traces from HbA1c values and quarterly 7-point capillary glucose profiles collected during the original DCCT. Glycemia risk index values were computed from 14 days of CGM data. Rates of retinopathy progression were calculated across deciles of GRI and HbA1c. Poisson regressions quantified the relationship between each metric and complication risk. Receiver operating characteristic (ROC) curves evaluated the ability of GRI and HbA1c to discriminate between the DCCT intensive and conventional treatment groups. RESULTS: Increased GRI was associated with higher rates of retinopathy progression (β=0.03, P < .001, D2=0.97). Comparable results were observed using HbA1c for retinopathy progression (β=0.47, P<.001, D2=0.93). The glycemia risk index also demonstrated separation between the intensive and conventional treatment groups, with an ROC area of 0.91 (95% CI, 0.90-0.93) compared to 0.88 (95% CI, 0.86-0.90) for HbA1c. CONCLUSIONS: Glycemia risk index demonstrated a significant association with retinopathy progression within the Virtual DCCT framework. The relationship observed between GRI and retinopathy was comparable to that obtained using HbA1c. These findings extend the validation of GRI beyond clinician-based assessment and support its potential utility as a CGM-derived metric for glycemic risk assessment and diabetes complications.