Abstract
Aim
To investigate the association between objective sleep parameters and glycaemic variability determined by continous glucose monitoring (CGM) among patients with type 2 diabetes, given the significant role of sleep in glycaemic control.
Methods
In this study, CGM was carried out in 28 patients with T2D (aged 62.3 ± 4.8 years, 57% women). Sleep characteristics were assessed by actigraphy within the CGM period. CGM‐derived outcomes included glucose level, and percentages of time in range (TIR) and time above range (TAR) during the monitoring period. Associations between intraindividual night‐to‐night variations in sleep characteristics and overall CGM outcomes were analysed using linear regression. Associations between sleep characteristics during each night and time‐matched CGM outcomes were analysed using linear mixed models.
Results
A total of 249 person‐days of CGM, coupled with 221 nights of sleep characteristics, were documented. Greater standard deviation (SD) of objective sleep duration (minutes) between measurement nights was associated with higher glucose level (coefficient 0.018 mmol/L [95% confidence interval {CI} 0.004, 0.033], P = 0.017), smaller proportion of TIR (% in observation period; coefficient −0.20% [95% CI −0.36, −0.03], P = 0.023), and greater proportion of TAR (coefficient 0.22% [95% CI 0.06, 0.39], P = 0.011). Later sleep midpoint (minutes from midnight) was associated with greater SD of glucose during the same sleep period (coefficient 0.002 minutes [95% CI 0.0001, 0.003], P = 0.037), longer nocturnal sleep duration was associated with smaller coefficient of variation of glucose level in the upcoming day (−0.015% [95% CI −0.03, −0.001], P = 0.041).
Conclusion
Objectively determined sleep duration and sleep midpoint, as well as their daily variability, are associated with CGM‐derived glucose profiles in T2D patients.
Keywords: chronotype, continuous glucose monitoring, daily variability, sleep, type 2 diabetes
1. INTRODUCTION
Despite the significant modulating effect of sleep and behavioural rhythms on glucose metabolism, it was not until recent years that their importance has been recognized in the clinical practice of type 2 diabetes (T2D). Evidence has tied insufficient sleep duration, and fragmented and disrupted sleep to poor glycaemic control among patients with T2D. 1 , 2 In addition, temporal misalignment between behavioural and the relatively stable endogenous circadian rhythms led by late chronotype, shift work and other factors related to alteration in sleep timing has similar detrimental influence on glycaemic control. 3 , 4 When left unattended, these problems may cause extra burden in the management of T2D. 5
It is of note that most studies associating sleep and glycaemic control were based on self‐reported sleep variables or average values of sleep variables derived from multiple‐night objective measurements (eg, wrist actigraphy), both of which reflect general sleep quality in a certain period. What has not been reported by these studies is night‐to‐night fluctuations in the sleep variables, which has been found to be more pronounced among individuals with a disrupted sleep and circadian pattern, and is highly correlated with glycaemic control. 6 , 7 , 8 , 9 Similar limitation applies to the observation of glycaemic outcomes, since commonly used clinical measurements such as fasting glucose and glycated haemoglobin (HbA1c) could not give a full picture of longitudinal glucose variations.
In providing real‐time glucose level readings, continuous glucose monitoring (CGM) is an emerging tool for optimizing long‐term glycaemic control in T2D. 10 , 11 Notably, recent studies have explored the link between objective sleep variables and CGM‐derived glycaemic control. 12 , 13 , 14 However, similar investigation has not been carried out among individuals with T2D. In the present study, multiple‐night objective sleep measurement by wrist actigraphy was performed during a 14‐day CGM period among patients with T2D. By hypothesizing that both nocturnal sleep and circadian misalignment have an impact on glycaemic control, we aimed to investigate whether sleep and circadian variables, as well as their night‐to‐night variations were associated with CGM‐derived glycaemic outcomes among T2D individuals in real‐life settings. We also sought to explore the potential imminent impact of nocturnal sleep on glucose variation among individuals with T2D.
2. METHODS
2.1. Study population
This was a community‐based observational study among middle‐aged and older adults resident in Nanjing Municipality, Jiangsu Province, China. Participants were recruited via online or printed advertisement delivered through the Maigaoqiao Community Health Service Center, Qixia District, Nanjing. Eligible participants were T2D patients aged 50 years or older, with a duration of T2D of at least 1 year. T2D was diagnosed by physicians according to the following criteria: random blood glucose ≥11.1 mmol/L; fasting blood glucose ≥7.0 mmol/L; 2‐hour post‐oral glucose tolerance test blood glucose ≥11.1 mmol/L; or HbA1c ≥ 48 mmol/mol (6.5%). Exclusion criteria included type 1 diabetes, present shift work, severe kidney or liver disease, diabetic ketoacidosis, recurrent hypoglycaemic events in the previous 3 months, dementia and other mental disorders. The study protocol was approved by the ethics committee of the Affiliated Rehabilitation Hospital of Nanjing Sport Institute in accordance with the principles of the Declaration of Helsinki. Participation was voluntary, and a written consent form was obtained from each participant prior to the measurements.
2.2. Continuous glucose monitoring
A CGM system (FreeStyle Libre Pro; Abbott Diabetes Care Ltd, Witney, UK) recorded participants' glucose values for 14 consecutive days in September 2020. Participants were asked to keep their normal routine for medication, eating and other daily activities during the CGM period. The CGM sensor for subcutaneous interstitial glucose monitoring was attached by a technician on the back of the upper arm. Sensors were attached between 7:00 am and 11:00 am on Day 1 and detached between 1:00 pm and 5:00 pm on Day 14. Average values for each 15‐minute ambulatory glucose level were automatically stored by the sensor. Following the monitoring period, glucose data were downloaded through the reader of the system and exported using FreeStyle Libre Pro desktop software. Participants did not have access to the CGM data during the monitoring period. Mean glucose values, standard deviation (SD) and coefficient of variation (CV; calculated as [SD/mean] × 100%) of the glucose values, time spent within target glucose range (TIR; 3.9 to 10 mmol/L), above target range (TAR; >10 mmol/L), and below target range (TBR; <3.9 mmol/L) were calculated based on 15‐minute intervals. 15
2.3. Objective assessment of sleep
Participants were instructed to wear an actigraphy device (GT3x; ActiGraph LLC, Pensacola, Florida) on their non‐dominant wrist for at least 7 consecutive days during the CGM period. 16 They were asked to keep wearing the device except for when showering, bathing and swimming. Actigraphy devices were handed to the participants on the date of CGM sensor attachment and collected after the end of the CGM period. A sampling frequency of 30 Hz was set for determining acceleration. Participants' usual bedtime and morning awakening time collected by questionnaire, as well as light intensity (in lux) determined by the actigraphy device were used to assist actigraphy data scoring. 17 The Cole‐Kripke algorithm was applied to identify sleep and wake state in 1‐minute epochs. 18 Total sleep time (TST), wake after sleep onset (WASO), number of awakenings, sleep efficiency (calculated as TST/TST + WASO), and sleep midpoint (mid timepoint between sleep onset and final morning awakening) of each night were obtained or calculated.
2.4. Confounding factors
Participants' age, gender, duration of diabetes, and current antidiabetic treatment regimen were ascertained by a physician. Weight, height and blood pressure were measured on the date of CGM sensor attachment at the community health service centre. Body mass index was calculated as weight (kg)/height (m)2.
2.5. Data analyses and statistics
A graphic interpretation of the study is shown in Figure 1. Analyses were based on days (midnight to midnight) having complete CGM and actigraphic sleep measurements. If wear time of the actigraphy device was not sufficient to estimate a complete nocturnal sleep period (determined as no valid wear time between 6:00 pm and 2:00 am, or between 4:00 am and 12:00 pm [midday]), the corresponding measurement days (115 person‐days) were excluded. 19 , 20 Furthermore, according to the instruction given by the CGM system manufacturer, the first 12‐hour glucose values following sensor attachment were not included in the analysis. No other CGM record within the period of interest was excluded or missing for the analysis. Mean values of sleep variables across the days of analysis for each participant were calculated. Intraindividual night‐to‐night sleep variability was determined by SDs and CVs of the sleep variables across these nights. Correlations between sleep variables or sleep regularity and CGM outcomes of the corresponding days were tested by linear regression analysis.
FIGURE 1.
Graphical interpretation of the study. CGM, continuous glucose monitoring
Actigraphic sleep‐wake state (per minute) and CGM‐derived ambulatory glucose level (per 15 minutes) were synchronized to distinguish between glycaemic outcomes in sleep and wake states. A sleep period was defined from time of sleep onset to time of final awakening in the next morning. A wake period was defined from time of morning awakening until time of next sleep onset in the evening hours, or until the time when the CGM sensor was detached (for the last measurement day). When a 15‐minute CGM interval included a sleep onset or final awakening timepoint, the longer side of sleep/wake state was chosen to label the interval. Nights with synchronized sleep and CGM data (221 person‐nights) were involved to test the association between nocturnal sleep variables and simultaneous CGM outcomes. In addition to CGM‐derived mean glucose, TIR, TAR and TBR, SD and CV of the ambulatory glucose values during the period of interest were calculated. Similarly, the relationship between nocturnal sleep variables and CGM outcomes in the next wake period were tested. Considering the repeated‐measurement nature of the values for each participant, linear mixed‐effects models were introduced to examine the associations between nocturnal sleep variables and simultaneous CGM outcomes or CGM outcomes in the next wake period, where assumptions were tested in a two‐level model (Level 1: within‐person; Level 2: between‐person). 21 All analyses were conducted using Stata v.17.0 (Stata LLC, College Station, Texas). P values <0.05 were taken to indicate statistical significance. Age, gender, body mass index, systolic blood pressure, duration of diabetes, and antidiabetic treatment (under treatment/no treatment) were adjusted in the linear regression and linear mixed‐model analyses.
2.6. Sensitivity analyses
Instead of using a continuous TST measurement, we dichotomized the mean TST for each individual and single‐night TST into normal (360‐480 minutes) and non‐recommended (<360 or > 480 minutes) groups. Analyses of covariance adjusted for age, gender, body mass index, systolic blood pressure, duration of diabetes, and antidiabetic treatment were performed to compare the CGM variables between participants with and without a normal mean TST during the observational period. Similar mixed‐effects models were introduced to assess the associations between single‐night TST groups and simultaneous CGM outcomes, as well as the CGM outcomes in the next wake period. Due to the sample size of the present study, we did not further distinguish between short and long sleep duration.
3. RESULTS
3.1. Participant characteristics
A total of 28 patients with T2D (16 women, 57.1%) participated in this study. Data from all participants were included in the analysis. The basic characteristics of the study population are given in Table 1. Participants were aged 62.3 ± 4.8 years, with a T2D duration of 8.1 ± 4.8 years. More than 85% of the participants (n = 24) were receiving antidiabetic pharmacotherapy. Metformin (n = 15, 53.6%), sulphonylureas (n = 11, 39.3%) and alpha‐glucosidase inhibitors (n = 5, 17.9%) were the most common oral antidiabetics administered in the population. Five participants (17.9%) were receiving insulin therapy.
TABLE 1.
Basic characteristics of the participants
| Characteristic | Mean ± SD/N (%) |
|---|---|
| Number of participants | 28 |
| Age, years | 62.3 ± 4.8 |
| Gender | |
| Female | 16 (57.1) |
| Male | 12 (42.9) |
| Body mass index, kg/m2 | 23.3 ± 3.0 |
| Duration of diabetes, years | 8.1 ± 5.3 |
| Systolic blood pressure, mmHg | 127.5 ± 14.0 |
| Diastolic blood pressure, mmHg | 79.6 ± 8.6 |
| Fasting glucose, mmol/L | 7.9 ± 1.5 |
| Fasting insulin, μIU/mL | 20.1 ± 40.4 |
| HbA1c, % | 7.0 ± 0.9 |
| Antidiabetic pharmacotherapy | |
| Yes | 24 (85.7) |
| No | 4 (14.3) |
| Type of antidiabetic drug | |
| Metformin | 15 (53.6) |
| Sulphonylureas | 11 (39.3) |
| Alpha‐glucosidase inhibitors | 5 (17.9) |
| DPP‐4 inhibitors | 2 (7.1) |
| Glitazone | 1 (3.6) |
| Insulin | 5 (17.9) |
| Other medications | |
| Antihypertensive | 14 (50.0) |
| Cholesterol‐lowering drug | 7 (25.0) |
| Antiplatelet | 4 (14.3) |
| Actigraphy sleep variables | |
| Night of measurement | 7.9 ± 2.5 |
| TST, min | 363 ± 74 |
| Sleep efficiency, % | 86.2 ± 5.8 |
| WASO, minutes | 56 ± 26 |
| Number of awakenings | 19 ± 8 |
| Sleep midpoint, hh:Mm a | 02:51 ± 77 |
| CGM variable | |
| Mean glucose, mmol/L | 7.6 ± 1.3 |
| SD of glucose, mmol/L | 2.2 ± 0.6 |
| CV of glucose, % | 29.3 ± 5.2 |
| TIR, % of total duration | 80.5 ± 14.0 |
| TAR, % of total duration | 17.4 ± 14.6 |
| TBR, % of total duration | 2.1 ± 3.8 |
Abbreviations: CGM, continuous glucose monitoring; CV, coefficient of variation; HbA1c, glycated haemoglobin; TIR, time in range; TAR, time above range; TBR, time below range; TST, total sleep time; SD, standard deviation; WASO, wake after sleep onset.
Data shown as time in 24 hours ± SD in minutes.
3.2. Sleep variables determined by wrist actigraphy
A total 221 person‐nights of valid actigraphy sleep data (7.9 ± 2.5 nights per participant) were collected. The mean value for nocturnal TST in all measurement nights was 363 ± 74 minutes, for WASO it was 56 ± 26 minutes, and for sleep efficiency it was 86.2 ± 5.8%. The mean sleep midpoint was at 02:51 AM ± 77 minutes. The mean values of sleep variables across measurement nights in each participant are listed in Table S1.
3.3. Association between night‐to‐night sleep regularity and CGM outcomes
During the period of interest, CGM was 100% active. The associations between intraindividual night‐to‐night variation of sleep variables and CGM glucose outcomes over the observation period are shown in Table 2. SD of TST (in minutes) between multiple‐night measurements was correlated with mean glucose value (in mmol/L; coefficient 0.018 [95% confidence interval {CI} 0.004, 0.033], P = 0.017), SD of glucose values (coefficient 0.008 [95% CI 0.001, 0.015], P = 0.035), and percentage of TAR spent in the period (coefficient 0.22 [95% CI 0.06, 0.40], P = 0.011), and was negatively correlated with percentage of TIR in the period (coefficient −0.20 [95% CI −0.36, −0.03], P = 0.023). These findings were consistent when using CV (in %) as the parameter for the magnitude of night‐to‐night variation (5.5 [95% CI 0.7, 10.4], P = 0.028, associated with mean glucose; 2.43 [95% CI 0.04, 4.81], P = 0.047, with % TIR in total duration; −58.1 [95% CI −113.0, −3.2], P = 0.039, with % TIR in total duration; 66.5 [95% CI 11.0, 121.9], with % TAR in total duration). A greater CV in sleep midpoint was associated with higher mean glucose (coefficient 4.2 [95% CI 1.2, 7.1], P = 0.008), lower percentage of TIR in the period (coefficient −43.7 [95% CI −77.2, −10.1], P = 0.013), and higher percentage of TAR in the period (coefficient 46.8 [95% CI 12.4, 81.1], P = 0.010). Mean values of sleep variables across multiple nights were not found to be associated with any CGM‐derived glucose metrics over the observation period (P > 0.05).
TABLE 2.
Association between sleep characteristics, sleep regularity and continuous glucose monitoring outcomes over the observation period
| Variable | Mean glucose, mmol/L | SD of glucose, mmol/L | CV of glucose, % | TIR, % of total duration | TAR, % of total duration | TBR, % of total duration |
|---|---|---|---|---|---|---|
| Coefficient (95% CI) | Coefficient (95% CI) | Coefficient (95% CI) | Coefficient (95% CI) | Coefficient (95% CI) | Coefficient (95% CI) | |
| TST, min | 0.001 (−0.01, 0.01) | −0.001 (−0.006, 0.004) | −0.02 (−0.06, 0.02) | 0.002 (−0.12, 0.12) | 0.004 (−0.12, 0.13) | −0.01 (−0.34, 0.23) |
| SD of TST, min | 0.018 (0.004, 0.033)* | 0.008 (0.001, 0.015)* | 0.03 (−0.04, 0.09) | −0.20 (−0.36, −0.03)* | 0.22 (0.06, 0.40)* | −0.03 (−0.07, 0.16) |
| CV of TST, % | 5.5 (0.7, 10.4)* | 2.43 (0.04, 4.81)* | 9.9 (−11.1, 30.9) | −58.1 (−113.0, −3.2)* | 66.5 (11.0, 121.9)* | −8.4 (−22.6, 5.8) |
| Sleep efficiency, % | 0.01 (−0.10, 0.11) | 0.01 (−0.04, 0.06) | 0.05 (−0.37, 0.46) | −0.26 (−1.43, 0.92) | 0.12 (−1.10, 1.35) | 0.13 (−0.15, 0.41) |
| SD of sleep efficiency, % | 0.03 (−0.14, 0.20) | 0.02 (−0.06, 0.10) | 0.18 (−0.48, 0.85) | −0.04 (−1.94, 1.87) | 0.31 (−1.66, 2.28) | −0.27 (−0.71, 0.17) |
| CV of sleep efficiency, % | 1.5 (−11.2, 14.3) | 1.2 (−4.9, 7.3) | 11.7 (−37.9, 61.4) | 4.3 (−137.9, 146.4) | 16.4 (−130.8, 163.7) | −20.7 (−53.5, 12.2) |
| WASO, min | −0.003 (−0.03, 0.02) | −0.003 (−0.02, 0.01) | −0.03 (−0.13, 0.06) | 0.09 (−0.19, 0.36) | −0.05 (−0.34, 0.24) | −0.03 (−0.10, 0.03) |
| SD of WASO, min | 0.01 (−0.03, 0.05) | 0.002 (−0.02, 0.02) | 0.01 (−0.16, 0.17) | −0.004 (−0.48, 0.47) | 0.08 (−0.41, 0.57) | −0.08 (−0.19, 0.03) |
| CV of WASO, % | 1.9 (−1.0, 4.8) | 0.6 (−0.8, 2.1) | 1.4 (−10.4, 13.3) | −16.0 (−48.8, 16.8) | 19.5 (−14.2, 53.2) | −3.5 (−11.4, 4.5) |
| Number of awakenings | −0.05 (−0.13, 0.04) | −0.02 (−0.07, 0.02) | −0.13 (−0.48, 0.23) | 0.53 (−0.46, 1.51) | −0.53 (−1.56, 0.49) | 0.01 (−0.24, 0.25) |
| Sleep midpoint, min from midnight | −0.001 (−0.01, 0.02) | 0.0003 (−0.004, 0.004) | 0.01 (−0.03, 0.04) | 0.02 (−0.08, 0.12) | −0.01 (−0.11, 0.09) | −0.01 (−0.03, 0.01) |
| SD of sleep midpoint, min | 0.032 (−0.001, 0.064) | 0.01 (−0.01, 0.03) | 0.02 (−0.12, 0.16) | −0.31 (−0.67, 0.06) | 0.35 (−0.02, 0.73) | −0.05 (−0.14, 0.04) |
| CV of sleep midpoint, % | 4.2 (1.2,7.1)* | 1.1 (−0.5, 2.7) | −0.5 (−14.3, 13.3) | −43.7 (−77.2, −10.1)* | 46.8 (12.4, 81.1)* | −3.1 (−12.5, 6.2) |
Note: Linear regression adjusted for age, gender, body mass index, systolic blood pressure, duration of diabetes, and antidiabetic treatment (binary, yes/no).
Abbreviations: CI, confidence interval; CV, coefficient of variation; SD, standard deviation; TAR, time above range; TBR, time below range; TIR, time in range; TST, total sleep time; WASO, wake after sleep onset.
P < 0.05.
3.4. Associations between nocturnal sleep variables and simultaneous CGM outcomes
By matching each measurement night's sleep variables with simultaneous CGM outcomes, we found that later sleep midpoint (in minutes from midnight) was associated with higher SD of overnight CGM glucose level (in mmol/L; coefficient 0.002 [95% CI 0.0001, 0.003], P = 0.037 [Table 3]). In addition, longer TST was associated with higher nocturnal glucose variation determined by both SD and CV of CGM glucose (0.002 [95% CI 0.001, 0.003], P < 0.001, with SD of glucose; 0.025 [95% CI 0.011, 0.038], P < 0.001, with CV of glucose).
TABLE 3.
Associations between nocturnal sleep variables and simultaneous continuous glucose monitoring outcomes
| Nocturnal variable | Mean glucose, mmol/L | SD of glucose, mmol/L | CV of glucose, % | % of TIR in TST | % of TAR in TST | % of TBR in TST |
|---|---|---|---|---|---|---|
| Coefficient (95% CI) | Coefficient (95% CI) | Coefficient (95% CI) | Coefficient (95% CI) | Coefficient (95% CI) | Coefficient (95% CI) | |
| TST, min | 0.0004 (−0.002, 0.003) | 0.002 (0.001, 0.003)* | 0.03 (0.01, 0.04)* | 0.01 (−0.04, 0.05) | −0.001 (−0.03, 0.03) | −0.004 (−0.04, 0.03) |
| Sleep efficiency, % | −0.01 (−0.04, 0.01) | 0.004 (−0.01, 0.02) | 0.09 (−0.05, 0.23) | −0.2 (−0.7, 0.2) | −0.03 (−0.31, 0.25) | 0.28 (−0.09, 0.64) |
| WASO, min | 0.003 (−0.003, 0.009) | −0.0003 (−0.003, 0.003) | −0.01 (−0.05, 0.02) | 0.07 (−0.03, 0.17) | −0.003 (−0.07, 0.06) | −0.06 (−0.15, 0.02) |
| Number of awakenings | 0.006 (−0.018, 0.030) | −0.003 (−0.014, 0.007) | −0.08 (−0.21, 0.05) | 0.003 (−0.001, 0.007) | −0.001 (−0.004, 0.001) | −0.002 (−0.006, 0.001) |
| Sleep midpoint, min from midnight | 0.003 (−0.001, 0.006) | 0.002 (0.0001, 0.003)* | 0.02 (−0.001, 0.03) | 0.02 (−0.03, 0.08) | 0.02 (−0.01, 0.06) | −0.04 (−0.09, 0.001) |
Note: Linear mixed model adjusted for age, gender, body mass index, systolic blood pressure, duration of diabetes, and antidiabetic treatment (binary, yes/no).
Abbreviations: CI, confidence interval; CV, coefficient of variation; SD, standard deviation; TAR, time above range; TBR, time below range; TIR, time in range; TST, total sleep time; WASO, wake after sleep onset.
P < 0.05.
3.5. Associations between nocturnal sleep variables and next‐day CGM outcomes
The links between nocturnal sleep variables and CGM outcomes the next day (determined as morning wake‐up time to next sleep‐onset time at night) are shown in Table 4. A longer TST was linked to a lower CV of glucose on the subsequent day (coefficient −0.015 [95% CI −0.030, −0.001], P = 0.041). No other associations were observed between nocturnal sleep variables and next‐day CGM outcomes.
TABLE 4.
Associations between nocturnal sleep variables and continuous glucose monitoring outcomes in the next wake period
| Variable | Mean glucose, mmol/L | SD of glucose, mmol/L | CV of glucose, % | % of TIR in wake period | % of TAR in wake period | % of TBR in wake period |
|---|---|---|---|---|---|---|
| Coefficient (95% CI) | Coefficient (95% CI) | Coefficient (95% CI) | Coefficient (95% CI) | Coefficient (95% CI) | Coefficient (95% CI) | |
| TST, min | 0.003 (−0.0002, 0.005) | −0.001 (−0.002, 0.001) | −0.015 (−0.03, −0.001)* | −0.03 (−0.07, 0.003) | 0.03 (−0.004, 0.07) | −0.001 (−0.01, 0.002) |
| Sleep efficiency, % | −0.02 (−0.04, 0.01) | −0.01 (−0.03, 0.001) | −0.07 (−0.21, 0.08) | 0.07 (−0.30, 0.44) | −0.09 (−0.45, 0.28) | 0.02 (−0.02, 0.06) |
| WASO, min | 0.004 (−0.002. 0.01) | 0.002 (−0.001, 0.01) | 0.002 (−0.03, 0.04) | −0.02 (−0.11, 0.07) | 0.02 (−0.06, 0.11) | −0.01 (−0.02, 0.002) |
| Number of awakenings | 0.01 (−0.01,0.04) | −0.0001 (−0.013, 0.013) | −0.07 (−0.19, 0.06) | −0.07 (−0.41, 0.27) | 0.10 (−0.24, 0.44) | −0.02 (−0.06, 0.01) |
| Sleep midpoint, min from 00:00 | −0.0004 (−0.004, 0.003) | −0.001 (−0.003, 0.001) | −0.01 (−0.03, 0.004) | 0.03 (−0.02, 0.08) | −0.03 (−0.08, 0.02) | −0.001 (−0.01, 0.003) |
Note: Linear mixed model adjusted for age, gender, body mass index, systolic blood pressure, duration of diabetes, and antidiabetic treatment (binary, yes/no).
Abbreviations: CI, confidence interval; CV, coefficient of variation; SD, standard deviation; TAR, time above range; TBR, time below range; TIR, time in range; TST, total sleep time; WASO, wake after sleep onset.
P < 0.05.
3.6. Results of sensitivity analyses
No differences between the normal and non‐recommended mean TST groups were found regarding the CGM outcomes in the observation period (Table S2). Replacing continuous TST with such dichotomized TST exposure in intraindividual measurements did not exhibit any significant link with the CGM outcomes during sleep or on the next day (Table S3).
4. DISCUSSION
In this study among 28 community‐dwelling middle and older‐aged T2D patients with an average disease course of 8.1 years, we observed a correlation between actigraphy‐determined night‐to‐night sleep regularity and glycaemic control over a 249‐person‐day CGM period. Specifically, a 1‐hour increment in the SD of nocturnal sleep duration across multiple nights was associated with an approximately 1.1‐mmol/L (95% CI 0.2‐2.0) higher mean glucose value, 12% 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 less TIR, and 13% 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 more TAR during the total CGM period. A further analysis suggested that delaying sleep midpoint by 1 extra hour was associated with a 0.12‐mmol/L (95% CI 0.01‐0.18) increased SD of the glucose level during the same sleep period. Moreover, 1 hour longer nocturnal sleep duration was associated with a 0.9% (95% CI 0.1‐1.8) smaller CV of glucose level during the next day, suggesting a more stable daytime glycaemic profile.
Sleep has long been associated with glucose metabolism. In individuals with T2D, mounting evidence supports that either short or long sleep duration is linked with poorer glycaemic management. An earlier meta‐analysis involving patients with T2D suggested that, compared with normal sleep duration (6–8 hours), short (<6 hours) and long (>8 hours) sleep durations were associated with a 0.23% and a 0.13% increased HbA1c, respectively. 22 A more recent study focusing on 962 untreated T2D patients and prediabetic individuals showed a similar association between short (<5 hours) or long (>8 hours) sleep duration and HbA1c, as well as the correlation between long sleep duration and higher morning fasting glucose. 23 However, it must be kept in mind that the aforementioned findings were based on self‐reported sleep duration, which could yield a significantly different estimate from the TST assessed by wrist actigraphy. 24 , 25 The discrepancy between self‐reported and actigraphic sleep duration can be magnified in individuals with poor overall health conditions and insomnia symptoms, both of which are common circumstances in patients with T2D. 25 , 26 In the present study, no direct association between actigraphy‐derived mean sleep duration and CGM outcomes over the observation period were noted in individuals with T2D. These negative findings warrant further verification in larger samples.
Notwithstanding the absence of a relationship between mean values of sleep variables and CGM‐derived glycaemic control, the variability of TST and sleep midpoint between nights were correlated with glycaemic control among patients with T2D. Our finding was consistent with a recent study investigating the associations between 7‐day actigraphy sleep parameters and HbA1c among 172 T2D patients from the Netherlands, where greater variability in sleep duration and sleep midpoint were both associated with higher HbA1c. 27 The study further suggested that, among different objective sleep variables including sleep duration and sleep efficiency, the variability of sleep duration between nights was most strongly correlated with HbA1c. Another recent study using an 8‐day sleep diary to record sleep variables also reported significant associations between sleep duration variability/sleep midpoint variability and HbA1c among 56 Chinese T2D patients. 28 The mechanism that links variability in sleep duration and midpoint with glycaemic control remains largely obscure. Noticeably, daily intraindividual variability of sleep has been found to be more pronounced among individuals with insomnia symptoms, poor subjective sleep quality, and late chronotype. 29 These phenotypes may impair glycaemic control through changes in daily behaviours, such as reduced level of physical activity and higher dietary energy density, 30 , 31 thus making glucose level more likely to exceed the upper limit of the recommended range in CGM. Further supporting the mechanism, recent evidence from 953 healthy adults showed that a person's deviation from their usual sleep pattern was associated with higher postprandial glucose level. 32
The present study was among the first to examine the relationships between objective sleep variables and continuous glucose outcomes among adults with T2D. Our data indicated that timing of nocturnal sleep was directly associated with glucose variability during the same night. Although we did not find a significant association between sleep midpoint and mean glucose value or percentage of TIR/TAR/TBR, attention must be paid to circadian‐related variation of glycaemia because greater glycaemic variability has been related to cardiovascular complications in T2D patients. 33 Preliminary evidence shows that severity of obstructive sleep apnoea, a possible trigger of nocturnal glycaemic fluctuations, has a U‐shaped association with chronotype. 34 , 35 Moreover, a later timing of food consumption in the nights with a postponed sleep midpoint might be another reason behind greater glucose variation during sleep. However, studies are warranted to further elucidate the above‐mentioned candidate mechanisms.
In addition to the immediate impact of sleep timing on nocturnal glycaemic control, our results suggest that, in T2D patients, a longer nocturnal sleep duration could predict a more stable glycaemic profile in the upcoming day. Extending sleep duration by 1.2 hours in overweight adults with a habitual sleep duration of less than 6.5 hours per night has been shown to reduce their daily calorie intake by 270 kcal, compared to controls without sleep extension. 36 We speculate that, in real‐life settings, sufficient sleep duration has an imminent impact on the calorie intake among T2D patients, thus lowering the risk of postprandial hyperglycaemia and reducing daytime glucose variability. Another possible explanation is that prolonging the sleep duration has a beneficial effect on insulin sensitivity, which has been exhibited among healthy adults having a similar habitual sleep duration to our participants. 37
By implementing concurrent sleep and glucose monitoring among T2D patients in real‐life settings, the present study enabled us to observe the relationship between sleep characteristics and glycaemic control with a more detailed scope. Nevertheless, several points should be considered before applying our findings in the management of T2D. First, further investigations are warranted to confirm whether our results can be generalized into T2D patients of other ethnic backgrounds and with a higher body mass index. In addition, the possible impact of other sleep phenotypes such as insomnia symptoms and sleep apnoea on glycaemic control cannot be ruled out, 5 and such information has not been assessed in our investigation. A larger sample size is also needed to further investigate possible interactions between sleep duration/midpoint and sleep disorders regarding GGM‐derived glycaemic control. Furthermore, since a sleep diary or other tools for collecting bedtime were not used, we were not able to investigate the possible association between CGM outcomes and other important sleep variables, such as sleep onset latency and time in bed, and the sleep efficiency values in the present study were likely overestimated. Despite these limitations, our study suggests a strong correlation between objectively determined sleep duration, sleep midpoint, and daily variabilities in these, with CGM‐derived glycaemic profiles among T2D patients. Minimizing night‐to‐night sleep and circadian variability, securing sufficient sleep duration, and promoting an early chronotype may help to optimize glycaemic control among patients with T2D.
AUTHOR CONTRIBUTIONS
Xiao Tan conceptualized the study. Yan Zhao, Lijun Wei and Baoyi Chen recruited the participants. Yuchan Zheng, Yixin Tian, Qian Yu and Lijun Qin conducted data collection and data input. Xiao Tan and Yan Zhao performed statistical analysis, interpreted the data, and wrote the original draft. Christian Benedict, Kai Xu and Biao Sun carried out a critical reading of the draft. All authors reviewed and approved the final version of the article prior to submission. Yan Zhao and Xiao Tan are guarantors of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
CONFLICT OF INTEREST
No conflict of interest to disclose.
PEER REVIEW
The peer review history for this article is available at https://publons.com/publon/10.1111/dom.14930.
Supporting information
Table S1. Supporting information.
ACKNOWLEDGMENTS
This research was supported by the Sport and Health Project, Jiangsu Collaborative Innovation Center (Y.Z., JSCIC‐GP21007), Swedish Medical Research Society (X.T. P18‐0084), Åke Wiberg Stiftelse (X.T., M18‐0169, M19‐0266), Fredrik and Ingrid Thuring Foundation (X.T., 2019‐00488), Key Research and Development Program in Government of Jiangsu Province (Y.Z., BE2022828), and Kinesiology Laboratory Research Project, Nanjing Sport Institute (Y.Z. SYS202105). The funders had no role in the design and conduct of the study, nor the decision to prepare and submit the manuscript for publication.
Zhao Y, Zheng Y, Tian Y, et al. Objective sleep characteristics and continuous glucose monitoring profiles of type 2 diabetes patients in real‐life settings. Diabetes Obes Metab. 2023;25(3):823‐831. doi: 10.1111/dom.14930
Funding information Åke Wiberg Stiftelse, Grant/Award Numbers: M18‐0169, M19‐0266; Fredrik och Ingrid Thurings Stiftelse, Grant/Award Number: 2019‐00488; Jiangsu Collaborative Innovation Center for Sports and Health Engineering, Grant/Award Number: JSCIC‐GP21007; Key Research and Development Program in Jiangsu Province, Grant/Award Number: BE2022828; Svenska Sällskapet för Medicinsk Forskning, Grant/Award Number: P18‐0084
Contributor Information
Yan Zhao, Email: 2000080042@nsi.edu.cn.
Xiao Tan, Email: xiao.tan@zju.edu.cn.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
REFERENCES
- 1. Ohkuma T, Fujii H, Iwase M, et al. Impact of sleep duration on obesity and the glycemic level in patients with type 2 diabetes: the Fukuoka diabetes registry. Diabetes Care. 2013;36(3):611‐617. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Zhu B, Hershberger PE, Kapella MC, Fritschi C. The relationship between sleep disturbance and glycaemic control in adults with type 2 diabetes: an integrative review. J Clin Nurs. 2017;26(23‐24):4053‐4064. doi: 10.1111/jocn.13899 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Reutrakul S, Hood MM, Crowley SJ, et al. Chronotype is independently associated with glycemic control in type 2 diabetes. Diabetes Care. 2013;36(9):2523‐2529. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Morris CJ, Yang JN, Garcia JI, et al. Endogenous circadian system and circadian misalignment impact glucose tolerance via separate mechanisms in humans. Proc Natl Acad Sci U S A. 2015;112(17):E2225‐E2234. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Tan X, van Egmond L, Chapman CD, Cedernaes J, Benedict C. Aiding sleep in type 2 diabetes: therapeutic considerations. Lancet Diabetes Endocrinol. 2018;6(1):60‐68. [DOI] [PubMed] [Google Scholar]
- 6. Buysse DJ, Cheng Y, Germain A, et al. Night‐to‐night sleep variability in older adults with and without chronic insomnia. Sleep Med. 2010;11(1):56‐64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Reutrakul S, Van Cauter E. Interactions between sleep, circadian function, and glucose metabolism: implications for risk and severity of diabetes. Ann N Y Acad Sci. 2014;1311:151‐173. [DOI] [PubMed] [Google Scholar]
- 8. Van Cauter E, Polonsky KS, Scheen AJ. Roles of circadian rhythmicity and sleep in human glucose regulation. Endocr Rev. 1997;18(5):716‐738. [DOI] [PubMed] [Google Scholar]
- 9. Chontong S, Saetung S, Reutrakul S. Higher sleep variability is associated with poorer glycaemic control in patients with type 1 diabetes. J Sleep Res. 2016;25(4):438‐444. [DOI] [PubMed] [Google Scholar]
- 10. Martens T, Beck RW, Bailey R, et al. Effect of continuous glucose monitoring on glycemic control in patients with type 2 diabetes treated with basal insulin: a randomized clinical trial. JAMA. 2021;325(22):2262‐2272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Aleppo G, Beck RW, Bailey R, et al. The effect of discontinuing continuous glucose monitoring in adults with type 2 diabetes treated with basal insulin. Diabetes Care. 2021;44(12):2729‐2737. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Brandt R, Park M, Wroblewski K, Quinn L, Tasali E, Cinar A. Sleep quality and glycaemic variability in a real‐life setting in adults with type 1 diabetes. Diabetologia. 2021;64(10):2159‐2169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Griggs S, Grey M, Strohl KP, et al. Variations in sleep characteristics and glucose regulation in Young adults with type 1 diabetes. J Clin Endocrinol Metab. 2022;107(3):e1085‐e1095. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Griggs S, Hickman RL, Strohl KP, Redeker NS, Crawford SL, Grey M. Sleep‐wake characteristics, daytime sleepiness, and glycemia in young adults with type 1 diabetes. J Clin Sleep Med. 2021;17(9):1865‐1874. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Battelino T, Danne T, Bergenstal RM, et al. Clinical targets for continuous glucose monitoring data interpretation: recommendations from the international consensus on time in range. Diabetes Care. 2019;42(8):1593‐1603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Smith MT, McCrae CS, Cheung J, et al. Use of Actigraphy for the evaluation of sleep disorders and circadian rhythm sleep‐wake disorders: an American Academy of sleep medicine clinical practice guideline. J Clin Sleep Med. 2018;14(7):1231‐1237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Fekedulegn D, Andrew ME, Shi M, Violanti JM, Knox S, Innes KE. Actigraphy‐based assessment of sleep parameters. Ann Work Exposures Health. 2020;64(4):350‐367. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Cole RJ, Kripke DF, Gruen W, Mullaney DJ, Gillin JC. Automatic sleep/wake identification from wrist activity. Sleep. 1992;15(5):461‐469. [DOI] [PubMed] [Google Scholar]
- 19. Choi L, Liu Z, Matthews CE, Buchowski MS. Validation of accelerometer wear and nonwear time classification algorithm. Med Sci Sports Exerc. 2011;43(2):357‐364. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Knaier R, Höchsmann C, Infanger D, Hinrichs T, Schmidt‐Trucksäss A. Validation of automatic wear‐time detection algorithms in a free‐living setting of wrist‐worn and hip‐worn ActiGraph GT3X. BMC Public Health. 2019;19(1):244. doi: 10.1186/s12889-019-6568-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Snijders TAB, Bosker RJ. Multilevel Analysis. 2nd ed. London: Sage Publications; 2012. [Google Scholar]
- 22. Lee SWH, Ng KY, Chin WK. The impact of sleep amount and sleep quality on glycemic control in type 2 diabetes: a systematic review and meta‐analysis. Sleep Med Rev. 2017;31:91‐101. [DOI] [PubMed] [Google Scholar]
- 23. Mokhlesi B, Temple KA, Tjaden AH, et al. Association of Self‐Reported Sleep and Circadian Measures with Glycemia in adults with prediabetes or recently diagnosed untreated type 2 diabetes. Diabetes Care. 2019;42(7):1326‐1332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Cespedes EM, Hu FB, Redline S, et al. Comparison of self‐reported sleep duration with Actigraphy: results from the Hispanic community health study/study of Latinos Sueño ancillary study. Am J Epidemiol. 2016;183(6):561‐573. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Matthews KA, Patel SR, Pantesco EJ, et al. Similarities and differences in estimates of sleep duration by polysomnography, actigraphy, diary, and self‐reported habitual sleep in a community sample. Sleep Health. 2018;4(1):96‐103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Scarlett S, Nolan HN, Kenny RA, O'Connell MDL. Discrepancies in self‐reported and actigraphy‐based sleep duration are associated with self‐reported insomnia symptoms in community‐dwelling older adults. Sleep Health. 2021;7(1):83‐92. [DOI] [PubMed] [Google Scholar]
- 27. Brouwer A, van Raalte DH, Rutters F, et al. Sleep and HbA1c in patients with type 2 diabetes: which sleep characteristics matter Most? Diabetes Care. 2020;43(1):235‐243. [DOI] [PubMed] [Google Scholar]
- 28. Zhu B, Kapella MC, Zhao X, Fritschi C. Intra‐individual variability in sleep is related to glycaemic control in adults with type 2 diabetes. J Adv Nurs. 2020;76(4):991‐998. [DOI] [PubMed] [Google Scholar]
- 29. Bei B, Wiley JF, Trinder J, Manber R. Beyond the mean: a systematic review on the correlates of daily intraindividual variability of sleep/wake patterns. Sleep Med Rev. 2016;28:108‐124. [DOI] [PubMed] [Google Scholar]
- 30. Tan X, Alén M, Cheng SM, et al. Associations of disordered sleep with body fat distribution, physical activity and diet among overweight middle‐aged men. J Sleep Res. 2015;24(4):414‐424. [DOI] [PubMed] [Google Scholar]
- 31. Zuraikat FM, St‐Onge MP, Makarem N, Boege HL, Xi H, Aggarwal B. Evening Chronotype is associated with poorer habitual diet in US women, with dietary energy density mediating a relation of Chronotype with cardiovascular health. J Nutr. 2021;151(5):1150‐1158. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Tsereteli N, Vallat R, Fernandez‐Tajes J, et al. Impact of insufficient sleep on dysregulated blood glucose control under standardised meal conditions. Diabetologia. 2022;65(2):356‐365. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Siegelaar SE, Holleman F, Hoekstra JB, DeVries JH. Glucose variability; does it matter? Endocr Rev. 2010;31(2):171‐182. [DOI] [PubMed] [Google Scholar]
- 34. Kim LJ, Coelho FM, Hirotsu C, Bittencourt L, Tufik S, Andersen ML. Is the chronotype associated with obstructive sleep apnea? Sleep Breath. 2015;19(2):645‐651. [DOI] [PubMed] [Google Scholar]
- 35. Khaire SS, Gada JV, Utpat KV, Shah N, Varthakavi PK, Bhagwat NM. A study of glycemic variability in patients with type 2 diabetes mellitus with obstructive sleep apnea syndrome using a continuous glucose monitoring system. Clin Diabetes Endocrinol. 2020;5(6):10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Tasali E, Wroblewski K, Kahn E, Kilkus J, Schoeller DA. Effect of sleep extension on objectively assessed energy intake among adults with overweight in real‐life settings: a randomized clinical trial. JAMA Intern Med. 2022;182(4):365‐374. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Leproult R, Deliens G, Gilson M, Peigneux P. Beneficial impact of sleep extension on fasting insulin sensitivity in adults with habitual sleep restriction. Sleep. 2015;38(5):707‐715. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1. Supporting information.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.