The Importance of Sleep for Diabetes Associated Tasks and Outcomes

Overview

Adolescents with type 1 diabetes (T1D) experience more disturbed sleep compared to their healthy peers, especially because they tend to spend less time in deep sleep, the most restoring part of sleep, potentially impacting diabetes management. Disturbed sleep may adversely affect diabetes management which requires day-to-day decision-making, emotional and behavioural regulation, attention, and planning. Despite a massive increase in new technology, more than 50% of adolescents do not reach their glycaemic target. Lack of sleep impairing diabetes management including blood glucose monitoring may play an important role in reaching the goal. For approximately 4000 children and adolescents in Denmark living with T1D, sleep disturbances may therefore account for short and long-term diabetes complications. Our overall aims are to investigate: (1) If and how glycaemic variability (GV) influences sleep quality and sleep stages and (2) if and how poor sleep quality influences time-in-range (TIR), time-above-range (TAR) and time-below-range (TBR) the following day.

Background: Sleep distrubances are increasing among adolescents in Denmark. In 2017, a large health survey reported that 14.9 % of the 16-24 year-olds ( n= 7,627) had sleep problems over the last two weeks-an alarmingly high increase of 33.3% since the first survey in 2010. Adolescents with type 1 diabetes (T1D) experience more disturbed sleep compared to their healthy peers, especially because they tend to spend less time in deep sleep, the most restoring part of sleep. Disturbed sleep may adversely affect diabetes management which requires daily blood glucose monitoring (BGM), day-to-day decision-making, emotional and behavioural regulation, attention and planning (behavioural pathway). Diabetes management plays an important role in reaching glycemic targets and the worsening glycemic outcome may impact sleep and start a vicious circle. Currently, \>50% of adolescents do not reach the glycemic target. Furthermore, the disturbance of circadian rhythmicity may in itself have metabolic consequences (metabolic pathway). Disturbed sleep may contribute to the 40% increased risk of psychiatric disease (anxiety, depression and eating disorders) associated with childhood diabetes. A link between psychiatric disease, stress and sleep has been confirmed. For the approximately 4000 children and adolescents in Denmark living with T1D, this means sleep problems may contribute negatively and increase the risk of short and long-term diabetes complications. Over the past decades, the use of technology such as insulin pumps and continuous glucose monitors (CGM) in T1D treatment has increased dramatically, but still far from all reaches glycemic targets. This may, in part, be explained by the vicious circle and the additional disturbances from alarms and difficulties in initiating sleep due to itching or irritation from having devices attached to the body. The more beneficial effect on sleep from advanced insulin pumps like hybrid closed loop systems with less glycemic variability and fewer alarms overnight, has yet to be proved in children and adolescents. The term sleep is defined as sleep quality and duration. Disturbed sleep is defined as too short or long duration for age based on guidelines from the American Academy of Sleep Medicine (AASM) and/or poor quality sleep. The ability to be more resilient, have better coping strategies and more self-compassion may impact your sleep through the influence on rumination, reducing sleep latency (period from going to bed to falling asleep) and duration of wake-up periods, why resilience, coping and compassion are potentially modifiable factors. Other factors like social activities and media uses are important factors associated with changes in circadian rhythmicity, sleep length, sleep disturbances and poorer daytime performance. Furthermore, poor glycemic control is associated with sleep disorders like obstructive sleep apnea. The golden standard for sleep assessment is polysomnography (PSG), but it is limited because it (1) typically provides only 1-2 nights of information about a child's sleep, (2) requires in some cases observation outside the natural environment in a sleep lab and (3) has significant financial costs. We have access to a new home sleep test with eye movement detection (HST REM+) from Somnomedics GmbH which is validated against PSG in adults. The HST REM+ is a simplified PSG which measures sleep architecture (sleep stages) with three electrodes in the forehead and one behind the ear. The HST REM+ can make measurements for up to 5 nights at home with a tablet showing how to apply the electrodes. The HST REM+ is combined with ActiGraph wGT3X-BT (AG), a small wristwatch measuring movements and validated against PSG in adolescents. With the majority of children being treated with continuous glucose monitoring (CGM) and insulin pumps, we have the unique opportunity to combine data from these different objective sleep measures and thorough diabetes management data from downloads from insulin pump and CGM. This opens for the opportunity to study 24/7 variation in diabetes management and glycemic outcome including device alarm frequency and relate this to the current and previous night sleep. Additionally, we are able to study how advanced insulin pumps including hybrids closed loop systems impact sleep parameters such as sleep duration, sleep stages, sleep efficiency and sleep latency. Our overall aims are to investigate: (1) If and how glycaemic variability influences sleep quality and sleep stages and (2) if and how poor sleep quality influences time-in-range (TIR), time-above-range (TAR) and time-below-range (TBR) the following day. Main hypothesis 1. Glycaemic variability is negatively associated with sleep quality and time spent in deep sleep (sleep stage N3+N4) 2. Sleep quality is positively associated with time in range (TIR) the following day Study design: Observational, cross-sectional study. Methods: Children and adolescents with T1D are recruited from the Diabetes Outpatient Clinic, Steno Diabetes Center Copenhagen (SDCC), providing approximately 750 of the 900 children and adolescents using CGM and insulin pumps. One or both parents are invited to participate as well as healthy siblings as controls. The families are recruited from a notice in the waiting room, from the SDCC webpage and asked to participate during out-patient visits, where written and oral information are given. Questionnaires and clinical parameters will be collected. One-week simultaneous observation at home with AG, a sleep diary (which they receive on their mobile device after each night), HST REM+ (5 days) and diabetes management (insulin pump and CGM) will be offered. Eligibility criteria: Participants are included if they are 6-17 years of age with \> 6 months T1D duration and are using an insulin pump and CGM (all types). Exclusion criteria are severe psychiatric diseases using ICD-10 codes (e.g. ADHD, autism spectrum disorder, eating disorders, retarded mental development), unwillingness to participate and inability to read and understand Danish or English. Primary Outcome WP1: Sleep quality is defined as % time in deep sleep stages (N3 and N4). Secondary outcomes: Deep sleep (minutes), REM sleep % and minutes, sleep quality (AG), TST (HST REM+ and AG), sleep score from the Scandinavian Sleep Questionnaire - Children and Adolescents sleep (SSQ-CA). Exposure: GV during the night. Additionally, as exploratory analysis: Diabetes-specific stressors e.g., alarms and fear of hypoglycemia, resilience, quality of life, strengths and difficulties and stress in the caregiver. Covariates: Diabetes duration, treatment modality, age and gender. Primary outcome WP2: Daytime TIR. Secondary outcomes: a) Glycaemic outcomes (Daytime TAR and TBR, Day + night TIR, TBR, TAR, HbA1c, Glycaemic risk index, mean glucose, GV and standard deviation (SD)); b) diabetes-management (use of bolus-guide including blood glucose and carbohydrate from pump download). Exposure: Sleep quality is defined as % time in deep sleep stages (N3 and N4). Additionally, sleep quality (SSQ-CA and AG), REM sleep (HST REM+). Covariates: Diabetes duration, treatment modality, age, gender. Statistics/power: Statistical methods will include descriptive statistics, T-test, Spearman's correlation, RASCH modelling of questionnaire data. All variables in the models will be tested for effect size using univariate analysis. Multiple regression models will include explanatory variables and covariates depending on effect size and importance. Inverse probability weighting will be used to examine selection bias due to non-participation Regarding hypothesis 1: To detect a change of 3 % in deep sleep (N3) with an SD of 7 %, we have a power of 80% if 43 participate in the study (significance level of 5%). Regarding hypothesis 2: To detect a change of 5 % in TIR with an SD of 12 %, we have a power of 80% if 45 participate in the study (significance level of 5%).