Pitfalls of intermittent continuous glucose monitoring in patients with steroid diabetes

Only a few studies of continuous glucose monitoring (CGM) in patients with steroid diabetes have been published. Therefore, we investigated all patients with type 2 diabetes (n = 121) and steroid diabetes (n = 40) who used the FreeStyle Libre Pro^® device (Abbott Japan) at Gunma University Hospital between 2017 and 2019. Glycated hemoglobin (HbA1c), mean sensor glucose (SG), and glucose management indicator values were similar in both groups. However, the indices for glycemic variabilities, expressed as standard deviations and percent coefficients of variation, were higher in patients with steroid diabetes than in those with type 2 diabetes. The associations between HbA1c, mean SG, and time in range (TIR) when glucose values were 70-180, <70, or >180 mg/dL were assessed using Pearson's product-moment correlation coefficient, which demonstrated good correlations in both patient groups. However, patients with steroid diabetes had a higher SG and lower TIR than did counterparts with type 2 diabetes who had similar HbA1c levels. To examine the effect of prednisolone on CGM data, we divided patients with steroid diabetes into 2 subgroups according to prednisolone dose (≤5 and >5 mg), and found that the dose of this steroid impacted the associations between HbA1c and CGM data, mean SG, and TIR. In summary, our data highlight the importance of cautiously interpreting CGM data and the optimal HbA1c level in patients with steroid diabetes to prevent diabetes-related complications. Further analyses using other CGM devices are necessary to further validate our findings.