A Comparison of HbA1c and Index 60 Thresholds for Identifying Autoantibody Positive Individuals at High Risk for Type 1 Diabetes (T1D)

Current ADA standards include HbA1c values 5.7%-6.4% as an indicator of prediabetes. However, there is little information with regard to HbA1c as a specific predictor of T1D. We thus compared HbA1c with Index 60, a T1D marker, for identifying individuals at high risk for T1D. Index 60 was derived from a proportional hazards regression model for T1D that included log fasting C-peptide, 60-minute glucose, and 60-minute C-peptide. Diabetes Prevention Trial-type 1 participants with Index 60<2.00 at baseline were followed for the first (incident) OGTT exceeding an Index 60≥2.00 threshold (n=85). Similarly, those with HbA1c values <6.0% at baseline were followed for the first (incident) HbA1c ≥6.0% (n=77). Those with both incident Index 60≥2.00 and incident HbA1c ≥6.0 were excluded. At the incident OGTTs, those with HbA1c ≥6.0 were older (19.2±12.2 years vs. 12.7±7.2 years; p<0.001). The cumulative incidence for T1D was higher (log rank p<0.001) after Index 60≥2.00 than after HbA1c ≥6.0 (3- and 4-year risks for HbA1c ≥6.0: 0.41 and 0.45; 28/77 diagnosed vs. 3- and 4-year risks for Index 60≥2.00: 0.78 and 0.92; 60/85 diagnosed). The hazard ratio for Index 60≥2.00 vs. HbA1c ≥6.0 (with age adjustment) was significant [HR: 1.96 (1.20, 3.16); p=0.006]. Intervals from incident Index 60≥2.00 and HbA1c ≥6.0 to diagnosis were 1.2±0.9 (0.1-3.8) years and 1.0±1.0 (0-4.7) years, respectively (p=0.50). The 59 individuals who had OGTTs on the date of incident HbA1c ≥6.0 had higher 30-0 minute C-peptide (with age adjustment) than those with incident Index 60≥2.00 (2.8±1.9 ng/ml vs. 1.6±1.0 ng/ml; p<0.001).

In conclusion, Index 60≥2.00 was superior to HbA1c ≥6.0 at identifying individuals at high risk for T1D. Moreover, those with incident HbA1c ≥6.0 were much older with higher C-peptide levels, suggesting heterogeneity. These findings add support for including C-peptide together with glucose in developing predictors and pre-diagnostic endpoints that are specific for T1D.