P-338: HbA1c overestimates glycemic burden compared to glucose management indicator in type 1 diabetes
S d S F Lemos, G L Torres, L P S Lima, J P A Freitas, N C Monteiro, M d L P Machado, L M Martins, B F d O Santos, T Trevisan, N O d SantanaAbstract
Introduction
Glycated hemoglobin (HbA1c) and glucose management indicator (GMI) derived from continuous glucose monitoring (CGM) estimate glycemic control through distinct mechanisms. Whether GMI can reliably substitute HbA1c in clinical decision-making for type 1 diabetes (T1D) remains uncertain.
Objectives
To evaluate the agreement between laboratory HbA1c and CGM-derived GMI at six months, and to explore the association between time in range (TIR) and HbA1c in individuals with T1D.
Methods
A 6-month longitudinal study enrolled 53 individuals with T1D from a public endocrinology clinic in Northeastern Brazil. All participants used an intermittently scanned CGM. HbA1c was measured at baseline and at six months. Agreement between HbA1c and GMI at six months was assessed using the Bland-Altman method, intraclass correlation coefficient (ICC), and Pearson correlation. The association between TIR and HbA1c was assessed by linear regression.
Results
A total of 32(60.4%) of participants were under 25 years of age. Baseline HbA1c was 9.06±1.86%, diabetes duration was 11±8.72 years, and GMI at six months was 7.64±1.04%. Agreement between HbA1c and GMI was poor: mean bias was +0.80% (HbA1c systematically higher than GMI), with proportional bias indicating greater discrepancy in participants with worse glycemic control. Only 27% showed agreement within 0.5%, while 32% had discrepancies above 1%. ICC was moderate (0.616), despite strong Pearson correlation (r=0.838). TIR and HbA1c showed strong negative correlation (r=-0.789; R2=0.62), with each 10% increase in TIR corresponding to a 0.8% reduction in HbA1c.
Conclusions
HbA1c and GMI are not interchangeable in this population. In most participants, HbA1c exceeded GMI, revealing a predominantly unidirectional discrepancy whose magnitude surpassed previously reported rates, consistent with the proportional bias observed and the high baseline glycemic burden of this cohort, with direct implications for therapeutic decision-making.