Background and Aims

Glucose management indicator (GMI) is an updated approach for estimating HbA1c from continuous glucose monitoring (CGM) data. GMI has been validated in patients with type 1 diabetes, however data in type 2 diabetes (T2D) is lacking.

Methods

112 paired laboratory A1c and GMI values were collected from 50 T2D patients on insulin and oral antidiabetic drugs. Patients had professional CGM (Medtronic Ipro2 with Enlite® sensor) for seven days with laboratory HbA1c determined in the same week. The agreement between laboratory HbA1c and sensor-derived GMI was examined using Bland-Altman plot with repeated subject measures.

Results

24% of GMI values were within 0.25% of laboratory HbA1c and 40% of GMI values were within 0.5% from laboratory HbA1c values. Overall, GMI was lower than laboratory HbA1c (mean bias -0.8%, 95% limits of agreement -2.4 to 0.9%).The bias was greater at higher laboratory HbA1c values (see Fig).

Fig Bland-Altman plot of difference between Glucose Management index (GMI) minus laboratory HbA1c versus laboratory HbA1c in type 2 diabetes patient.

Conclusions

There are differences between sensor-derived GMI and laboratory HbA1c based on retrospective CGM in T2D patients, with GMI tending to be lower. This could be attributed to factors such as CGM duration, sensor accuracy, differences in glycation, acute change in glucose or patient behaviours on CGM. Clinicians should consider these factors when integrating both sets of values in therapeutic decision making.