Type 1 diabetes mellitus and its glycemic control affect sleep

Materials and methods Eighteen non-obese adults with T1D (8 males, age=26.3±5.1), without chronic complications were studied. The following instruments were used during ten consecutive days: sleep diaries (with Visual Analogue Scales for rating sleep quality), actimeters, and a home fingertip glucometer (6.41±1.5 tests a day). The analysis was made using fitted inflated beta regression models with GAMLSS for the response variable sleep quality rate (SQR). Subject-level covariables in this dataset included: number of hyperglycemia and hypoglycemia episodes in the previous day, number of awakenings during the night, morning awakening type (spontaneous or not), sleep rate (total sleep time in min divided by 1440), and the corresponding interactions. Results The SQR was negatively associated with the sleep rate (p<0.001), and positively associated with spontaneous awakening (p=0.042). The relationships between the other variables with SQR were the following: the SQR rises with the increase of both, hypoglycemic events and sleep rate (p=0.002). On the other hand, when the hypoglycemic or hyperglycemic events are associated with non-spontaneous awakening, the SQR decreases (p=0.02 for both glycemic conditions). However, the SQR increases when both the number of hyperglycemia and the number of night awakenings increase (p=0.01).