Association of the Glycemia Risk Index With Glycemic Metrics and Sensor Usage in a Real-World Pediatric Population...

The Glycemia Risk Index (GRI) is a recently developed composite measure designed to consolidate overall glycemic control into a single, interpretable score. The aim of this study was to investigate the associations between the GRI and glycemic metrics derived from continuous glucose monitoring (CGM), including sensor usage time, in children with diabetes using the FreeStyle Libre 2 Plus® CGM system (Abbott Diabetes Care, Witney, UK).

De-identified CGM data from 147 pediatric patients with diabetes in Saudi Arabia, treated at two governmental hospitals between January 2023 and September 2024, were analyzed. Glycemic metrics from the ambulatory glucose profile were recorded, and the GRI and its hypoglycemia and hyperglycemia components were calculated. Correlations between the GRI, its components, and time in range with glycemic metrics were assessed using Pearson and Spearman correlation coefficients. The associations between GRI and its components and two sensor usage time groups (70%-89% versus >90%) were evaluated using the independent t-test and Mann-Whitney U test. A p-value < 0.05 was considered statistically significant.

There was a negative correlation between GRI and time in range (r = -0.941), level one time above range (181-250 mg/dl; r = -0.447), time below range (< 70 mg/dl; r = -0.244), level one time below range (69-54 mg/dl; r = -0.248), and hypoglycemia component (r = -0.243) (all p-values <0.05). There was a positive correlation between GRI and average blood glucose (r = 0.923), glucose management indicator (r = 0.922), time above range (>180 mg/dl; r = 0.893), level two time above range (>250 mg/dl; r = 0.944), and hyperglycemia component (r = 0.929) (all p-values <0.05). Participants with lower sensor usage (70-89%) had significantly higher GRI values (median: 96.00; IQR: 75.80-100.00) compared to those with ≥90% usage (median: 82.40; IQR: 60.40-100.00; p = 0.004). The hyperglycemia component was also significantly higher in the lower usage group (mean: 53.65 vs. 46.62, p = 0.032).

Average glucose, glucose management indicator, and time in range showed negative correlations with GRI, while extreme hyperglycemia correlated positively. These results support GRI's role in assessing hyperglycemic exposure and treatment efficacy. Unexpected negative correlations with mild hyperglycemia and mild hypoglycemia warrant further studies. GRI and its hyperglycemic component improved with increased sensor usage, suggesting better glycemic control with higher CGM adherence.