The effect of sleep quality on glucose variability among adolescents with type 1 diabetes in China: A multi-central temporal longitudinal association analysis

Abstract

Study objectives

High glucose variability is common among adolescents with type 1 diabetes mellitus (T1DM). On person level, poor sleep quality has been reported to be an important factor associated with high glucose variability among adolescents with T1DM. However, on day level, the effect of sleep quality on glucose variability remains unclear, limiting temporal adjustment of treatment regimens. This study aimed to concurrently explore associations between sleep quality and glucose variability at both day and person levels among Chinese adolescents with T1DM based on the 24-h Recursive Cycle model.

Methods

A multi-center, seven-day, temporal longitudinal study was conducted among Chinese adolescents with T1DM. Glucose variability measures were calculated by fingertip blood glucose level at least seven times a day. Subjective sleep quality was measured by total sleep time, wake after sleep onset, number of awakenings, a score of sleep quality, sleep time, and wake time using a sleep diary. Objective sleep quality was assessed by Fitbit Inspire HR and included total sleep time, wake after sleep onset, and number of awakenings, rapid eye movement, light sleep time, deep sleep time, sleep time, wake time, sleep midpoint, and sleep efficiency. A multilevel linear regression model was performed to examine the associations between objective and subjective sleep quality and glucose variability at day and person levels. Gender, age, diabetes duration, complications documented within the preceding 6 months, HbA1c, and insulin pump therapy were controlled at person-level model.

Results

A total of 51 adolescents with T1DM participated in this study, which included 357 records of data. Only 21.57 % (N = 12) of adolescents met the recommended sleep time of 480 min per night measured by Fitbit. About a quarter (N = 11) of adolescents had a coefficient of variation of blood glucose >36 %. At person level, there was no significant association between sleep quality and glucose variability (p > 0.05). Multilevel models found significant associations between sleep quality and glucose variability at day level. Lower score of subjective sleep quality was significantly associated with higher standard deviation of blood glucose (p < 0.05) in the next day. Less Fitbit-measured light sleep time was significantly associated with higher standard deviation of blood glucose, and postprandial of glycemic excursions in the next day (p < 0.05). More Fitbit-measured awakenings and less rapid eye movement were associated with higher postprandial glycemic excursions in the next day (p < 0.05).

Conclusion

Nearly 80 % of Chinese adolescents with T1DM did not meet the recommended amount of sleep for their age group. They experienced more wakes after sleep onset at night and poorer sleep quality than their subjective experience. According to the findings on temporal relationships, intervention components targeting reducing wakes after sleep onset and improving subjective sleep quality would reduce glucose variability over time among Chinese adolescents with T1DM.