Prediction of nocturnal hypoglycemia unawareness by fasting glucose levels or post-breakfast glucose fluctuations in patients with type 1 diabetes receiving insulin degludec: A pilot study.

Abstract

To evaluate whether nocturnal asymptomatic hypoglycemia (NAH) can be predicted by fasting glucose levels or post-breakfast glucose fluctuations in patients with type 1 diabetes (T1D) receiving insulin degludec. Patients with T1D receiving insulin degludec underwent at-home CGM assessments. Indices for glycemic variability before and after breakfast included fasting glucose levels and the range of post-breakfast glucose elevation. For comparison, the patients were classified into those with NAH and those without. The optimal cut-off values for the relevant parameters were determined to predict NAH using ROC analysis. The study included a total of 31 patients (mean HbA1c values, 7.8 ± 0.7%), and 16 patients (52%) had NAH. Those with NAH had significantly lower fasting glucose levels than did those without (82 ± 48 mg/dL vs. 144 ± 69 mg/dL; P = 0.009). The change from pre- to post-breakfast glucose levels was significantly greater among those with NAH (postprandial 1-h, P = 0.028; postprandial 2-h, P = 0.028). The cut-off values for prediction of NAH were as follows: fasting glucose level <84 mg/dL (sensitivity 0.80/specificity 0.75/AUC 0.80; P = 0.004), 1-h postprandial elevation >69 mg/dL (0.75/0.67/0.73; P = 0.033), and 2-h postprandial elevation >99 mg/dL (0.69/0.67/0.71; P = 0.044). The results suggest that fasting glucose level of < 84 mg/dL had approximately 80% probability of predicting the occurrence of NAH in T1D receiving insulin degludec. It was also shown that the occurrence of hypoglycemia led to greater post-breakfast glucose fluctuations and steeper post-breakfast glucose gradients.