Abstract
Background
Sleep characteristics and glucose levels likely have a bidirectional relationship in type 1 diabetes (T1D), where highly variable glucose levels may impact habitual sleep timing, duration, and quality, and a disrupted or inconsistent sleep-wake cycle may negatively impact the ability to achieve recommended glucose levels. Adolescents and young adults (AYAs) with T1D are required to perform frequent blood glucose checks (SMBG) or use Continuous Glucose Monitoring (CGM) technology to self-monitor their glucose levels. However, evidence supporting the use of intermittently scanned CGM (isCGM) in AYAs is limited, and further research exploring the longer-term impacts on glucose testing and sleep patterns is required.
Aims
To describe habitual sleep-wake timing, duration, and quality in AYAs with T1D and above- target HbA1c; to investigate the impact of 6 months of use of isCGM compared to SMBG on sleep characteristics in AYAs with T1D; and to examine associations between the longer- term use of isCGM and glucose test frequency, haemoglobin A1c (HbA1c) and glucose time- in-range (TIR) among this population group.
Methods
Study 1 (an observational study) compares sleep characteristics in 64 young people (aged 13- 20 years) with T1D to 266 young people (aged 13-17 years) without diabetes. Sleep was assessed in both groups using seven days of Actigraphy to objectively measure sleep-wake timing and the Pittsburgh Sleep Quality Index (PSQI) questionnaire to subjectively measure sleep timing and assess sleep quality. Study 2 (a randomised controlled trial) assessed Actigraphy and PSQI outcomes to investigate the impact of six months of use of isCGM on sleep characteristics in AYAs with T1D and above-target HbA1c. In Study 3 (an observational study), the 6-month post-RCT extension phase data was analysed to determine the impacts of a further six months of use of isCGM on HbA1c and glucose TIR. Regression analyses were used to model between-group comparisons of sleep characteristics and glycaemic outcomes from baseline to 12 months after adjusting for various confounding factors.
ii
Results
In Study 1, AYAs with T1D reported later bedtimes (+36 minutes; p <0.05) and shorter sleep duration (-53 minutes; p <0.05) than controls without diabetes and were more likely to rate subjective sleep duration, efficiency, and quality as ‘poor’ (p <0.05); however, objectively measured sleep patterns were similar for both groups. In Study 2, the 6-month RCT showed objectively measured sleep-wake timing, and subjectively measured sleep quality was similar in the isCGM intervention and SMBG control groups. An analysis of the 6-month post-RCT extension phase outcomes (Study 3) revealed a mean difference in HbA1c from baseline of -4 mmol/mol [-0.4%]; p = 0.14) in the isCGM intervention group and -7 mmol/mol [-0.7%]; p = 0.08) in the SMBG control group at 12 months. The isCGM intervention group's mean rate of daily glucose testing was 2.4 times baseline rates at nine months (p < 0.001) but returned to baseline by 12 months (p = 0.091).
Conclusions
AYAs with T1D and above-target HbA1c demonstrate highly variable sleep-wake timing, with many not achieving age-based recommendations for hours of sleep each night. Despite a lack of difference in objectively measured sleep, more AYAs with T1D perceived their sleep to be of poor quality compared to their peers without diabetes. Overall, six months of use of isCGM in AYAs with T1D and above-target HbA1c did not impact objectively or subjectively measured sleep characteristics. The use of isCGM in addition to SMBG resulted in more daily glucose checks and lower HbA1c by nine months; however, these improvements were not sustained out to 12 months.