Abstract
Background
Since there are limited studies on the associations between glycemic variability (GV) and sleep quality or physical activity in subjects without diabetes, we evaluated the associations between GV, as assessed by continuous glucose monitoring (CGM), and both sleep quality and daily steps using wearable devices in healthy individuals.
Methods
Forty participants without diabetes were monitored by both an intermittently scanned CGM and a smartwatch-type activity tracker for 2 weeks. The standard deviation (SD) and coefficient of variation (CV) of glucose were evaluated as indices of GV. The activity tracker was used to calculate each participant's average step count per day. We also calculated sleep duration, sleep efficiency, and sleep latency based on data from the activity tracker. Spearman's correlation coefficient was used to assess the association between GV and sleep indices or daily steps. For each participant, periods were divided into quartiles according to step counts throughout the day. We compared mean parameter differences between the periods of lowest quartile and highest quartile (lower 25% and upper 25%).
Results
SD glucose was significantly positively correlated with sleep latency (R = 0.23, P < 0.05). There were no significant correlations among other indices in GV and sleep quality (P > 0.05). SD glucose and CV glucose levels in the upper 25% period of daily steps were lower than those in the lower 25% period in each participant (both, P < 0.01).
Conclusion
In subjects without diabetes, GV evaluated by intermittently scanned CGM was positively associated with the time to fall asleep. Furthermore, GV in the days of larger daily steps was decreased compared to the days of smaller daily steps in each participant.
Keywords: Glycemic variability, Wearable device, Intermittently scanned continuous glucose monitoring, Sleep quality
1. Background
Glycemic variability (GV) reflects fluctuations in blood glucose levels over time. Since it has been reported positive correlations between GV assessed by continuous glucose monitoring (CGM) and an oxidative stress marker [1] or coronary plaque rupture [2], GV is suggested to increase the development of cardiovascular complications [3]. Moreover, GV has also been associated with sleep quality [4,5] and physical activity [6]. It has been reported that poor sleep quality related with increased GV in patients with type 1 diabetes [4] and type 2 diabetes [5]. The increase of physical activity has been reviewed to decrease GV in patients with diabetes [6]. Most published data have been derived from patients with diabetes; there is only a limited body of research involving healthy subjects without diabetes.
Therefore, in the present subpopulation analysis a trial called “Assessment of glycemic variability and lifestyle behaviors in healthy nondiabetic individuals according to the categories of body mass index” [7], we evaluated the associations between GV, as assessed by CGM, and both sleep indices and daily steps using wearable devices in subjects without diabetes.
2. Methods
This study was carried out with the approval of the Keio University Institutional Review Board for Clinical Research (20211103) and registered as trial number UMIN000046858. Before enrollment, all participants provided written informed consent to the investigators and conformed the provisions of the Declaration of Helsinki.
2.1. Study design and participants
The trial examined the relationships between activities of daily living, GV assessed by wearable devices, and psychological state obtained from questionnaires. The trial design, as well as the inclusion and exclusion criteria for participants, have been described in detail elsewhere [7]. In the trial, participants were monitored simultaneously by wearing both an intermittently scanned CGM and a smartwatch-type activity tracker for 2 weeks. They were also evaluated by wearing both devices in the same way for 2 weeks after receiving dietary advice. A total of 40 office workers without diabetes participated in the present study.
2.2. Continuous glucose monitoring
CGM was performed twice for each participant using Freestyle Libre (Abbott Laboratories, Tokyo, Japan). The glucose profile data were downloaded, processed, visualized, and archived using licensed software. Data obtained from participants who underwent intermittently scanned CGM for more than 60% of the 2-week period were included in this analysis. The standard deviation (SD) and coefficient of variation (CV) of glucose were evaluated as indices of GV. We also assessed the time in range (TIR) and time in tight range (TITR), which represents the percentage of time spent within the target range of 70–180 or 70–140 mg/dL, respectively. TIR is not a GV index, but GV and TIR correlated with each other [8]. TITR describes time in normoglycaemia, and reported to correlate with TIR [9]. The definitions or interpretations of GV indices were described in a previous study [10].
2.3. Assessment of sleep indices and daily steps
Participants were instructed to wear a Fitbit Inspire 2 continuously for 2 weeks, with the option to remove it only for bathing (Fitbit Inc., Tokyo, Japan). The average step count per day, as recorded by the Fitbit Inspire 2, was calculated for each participant. Moreover, the Fitbit Inspire 2 provides data on time lying on the bed (min), sleep duration (min), and sleep state. We calculated sleep efficiency (%), defined as sleep duration/time lying on the bed and multiplied by 100, as well as sleep latency (min), which represents the time taken to fall asleep beyond the first 15 min after the beginning of lying on the bed.
2.4. Statistical analysis
Spearman's correlation coefficient was used to assess the association between two variables. For each participant, periods were divided into quartiles according to step counts throughout the day, and we compared mean parameter differences between the periods of lowest quartile and highest quartile (lower 25% and upper 25%) (Supplementary textbox1). We also evaluated daily CGM indices from 0 h to 24 h and sleep indices during the night in each period. The paired t-test was employed to analyze mean parameter differences between two groups of the lower 25% and upper 25% periods. All statistical analyses were performed using Python version 3.8.7. The level of significance was set at 5% (P < 0.05). Data are presented as mean ± SD.
3. Results
3.1. Associations between CGM indices and sleep indices
The mean age, body mass index (BMI), fasting plasma glucose and HbA1c of 40 participants, including 20 males, were 40 ± 12 years, 21.3 ± 2.5 kg/m2, 90.4 ± 11.5 mg/dl and 5.3 ± 0.4 %. Supplementary Table 1 presents the characteristics of participants with respect to metabolic, CGM indices and sleep indices. Table 1 displays the associations between CGM indices and sleep indices or mean daily steps. As a result, SD glucose was significantly positive correlated with sleep latency (R = 0.23, P < 0.05). There were no significant correlations among other indices (Table 1, P > 0.05).
Table 1.
Associations between CGM indices and sleep indices or mean daily steps.
| Correlation coefficient (R) | |||||
|---|---|---|---|---|---|
| Mean glucose (mg/dl) | SD glucose (mg/dl) | CV glucose (mg/dl) | Time in range (%) | Time in tight range (%) | |
| Time lying on the bed (min) | −0.117 | 0.013 | 0.049 | −0.016 | 0.135 |
| Sleep duration (min) | −0.164 | −0.011 | 0.044 | −0.010 | 0.168 |
| Sleep efficiency (%) | −0.155 | −0.101 | −0.042 | 0.041 | 0.116 |
| Sleep latency (min) | 0.022 | 0.233* | 0.210 | −0.207 | −0.175 |
| Mean daily steps | −0.052 | 0.099 | 0.139 | −0.121 | 0.003 |
CGM, continuous glucose monitoring; SD, standard deviation; CV, coefficient of variation. *P < 0.05.
3.2. Differences in CGM indices and sleep indices between the upper 25% and lower 25% periods of daily steps for each participant
The mean daily step count in the lower 25% period was 4670 ± 1981 steps, and that of the upper 25% was 15114 ± 3452 steps (Table 2). SD glucose and CV glucose levels were lower in the upper 25% period than in the lower 25% period (both, P < 0.01). Furthermore, TITR during the upper 25% period of daily steps showed an increase compared with those during the lower 25% period (87.5 ± 7.2% vs. 90.5 ± 8.6%, P < 0.01). There were no significant differences in mean glucose, TIR, or sleep indices between two groups.
Table 2.
Differences in CGM indices and sleep indices between the upper 25% and lower 25% periods of daily steps for each participant.
| Parameter | Lower 25% | Upper 25% | P value |
|---|---|---|---|
| Mean daily steps | 4670 ± 1981 | 15114 ± 3452 | <0.001 |
| CGM indices | |||
| Mean glucose (mg/dl) | 109.7 ± 9.1 | 108.2 ± 9.5 | 0.14 |
| SD glucose (mg/dl) | 20.0 ± 4.7 | 17.6 ± 4.5 | <0.001 |
| CV glucose (%) | 18.3 ± 4.1 | 16.4 ± 4.2 | <0.001 |
| Time in range (%) | 97.3 ± 3.3 | 97.7 ± 3.8 | 0.38 |
| Time in tight range (%) | 87.5 ± 7.2 | 90.5 ± 8.6 | <0.001 |
| Sleep indices | |||
| Time lying on the bed (min) | 429 ± 52 | 430 ± 51 | 0.95 |
| Sleep duration (min) | 373 ± 42 | 374 ± 45 | 0.94 |
| Sleep efficiency (%) | 87.0 ± 2.5 | 87.0 ± 2.4 | 0.89 |
| Sleep latency (min) | 7.7 ± 4.7 | 6.9 ± 3.8 | 0.30 |
CGM, continuous glucose monitoring; SD, standard deviation; CV, coefficient of variation.
4. Discussion
The increase in GV evaluated by CGM in patients with both type 1 and type 2 diabetes is reportedly associated with poor sleep quality [4,5]. The present study showed an association between SD glucose and sleep latency in subjects without diabetes, suggesting that even mild GV compared to patients with diabetes is also associated with sleep quality. Although poor sleep quality has been shown to induce the sympathetic nervous system and lead to insulin resistance and aberrant glucagon secretion [11,12], it has been unclear whether sleep quality affects GV in subjects without diabetes. Moreover, an increased time to fall asleep does not indicate poor sleep quality, and GV was not associated with other sleep indices, such as sleep duration or sleep efficiency, in our results. Therefore, we cannot judge the association between increased GV and poor sleep quality for subjects without diabetes in this study. Further investigations with a larger number of participants and better methods for evaluating sleep quality are necessary.
The present study observed improvement in GV during the upper 25% period of daily steps compared to the lower 25% period in each participant. Interestingly, there was no significant change in mean glucose between the two periods. These results suggested that daily walking depresses GV, but not mean glucose, in subjects without diabetes. A review found that physical activity and exercise have been reported to improve GV in patients with diabetes [6], but research in subjects without diabetes has been limited. A randomized crossover trial has shown that glucose CV and SD, as evaluated using CGM, were lower post-exercise compared to pre-exercise in subjects without diabetes [13]. However, in a cross-sectional study, physical activity status assessed through questionnaires did not show a significant relationship with GV indices [14]. Since there were also no relationships between GV and daily step count in the cross-sectional analysis of this study (Table 1), comparison between individuals might be a better method to evaluate the change in GV according to physical activity in subjects without diabetes.
The strength of this study is that it evaluated the associations between GV, as assessed by intermittently scanned CGM, and both sleep indices and daily steps using wearable devices in healthy individuals without diabetes. However, this study has some limitations. First, the statistical power to compare GV with sleep indices might be limited by the small sample size of participants. Second, the fluctuation with the mean absolute relative difference (MARD) of Freestyle Libre might have affected mild GV in this trial compared to patients with diabetes. Third, since this trial recruited healthy individuals from a single company, the results may not be generalizable to individuals without diabetes. Fourth, sleep quality was evaluated by only wearing devises in this analysis.
5. Conclusions
In subjects without diabetes, GV evaluated by intermittently scanned CGM was positively associated with the time to fall asleep. Furthermore, GV in the days of larger daily steps was decreased compared to the days of smaller daily steps in each participant.
Conflicts of interest
S.K., K.K., and T.K. belong to endowed chairs funded by Mori Building Co., Ltd.
CRediT authorship contribution statement
Jun Inaishi: Formal analysis, Methodology, Writing – original draft. Kazuhiro Kashiwagi: Conceptualization, Investigation, Formal analysis, Methodology, Supervision, Writing – original draft, Writing – review & editing. Shotaro Kinoshita: Conceptualization, Data curation, Investigation, Writing – review & editing. Yasuyo Wada: Investigation, Writing – review & editing. Sayaka Hanashiro: Data curation, Investigation. Kiko Shiga: Data curation, Investigation. Momoko Kitazawa: Data curation, Investigation. Shiori Tsutsumi: Data curation, Investigation. Hiroyuki Yamakawa: Investigation. Taishiro Kishimoto: Conceptualization, Supervision, Writing – review & editing.
Acknowledgements
We thank all participants who made the study possible and members of the Wellness Promotion Department at Mori Building Co., Ltd., who helped recruit the participants.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.metop.2023.100263.
Appendix A. Supplementary data
The following is the supplementary data to this article:
References
- 1.Monnier L., Mas E., Ginet C., et al. Activation of oxidative stress by acute glucose fluctuations compared with sustained chronic hyperglycemia in patients with type 2 diabetes. JAMA. 2006;295(14):1681–1687. doi: 10.1001/jama.295.14.1681. [DOI] [PubMed] [Google Scholar]
- 2.Teraguchi I., Imanishi T., Ozaki Y., et al. Impact of glucose fluctuation and monocyte subsets on coronary plaque rupture. Nutr Metabol Cardiovasc Dis. 2014;24(3):309–314. doi: 10.1016/j.numecd.2013.08.010. [DOI] [PubMed] [Google Scholar]
- 3.Saisho Y. Glycemic variability and oxidative stress: a link between diabetes and cardiovascular disease? Int J Mol Sci. 2014;15(10):18381–18406. doi: 10.3390/ijms151018381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Brandt R., Park M., Wroblewski K., et al. Sleep quality and glycaemic variability in a real-life setting in adults with type 1 diabetes. Diabetologia. 2021;64(10):2159–2169. doi: 10.1007/s00125-021-05500-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Yang Y., Zhao L.H., Li D.D., et al. Association of sleep quality with glycemic variability assessed by flash glucose monitoring in patients with type 2 diabetes. Diabetol Metab Syndrome. 2021;13(1):102. doi: 10.1186/s13098-021-00720-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Sparks J.R., Kishman E.E., Sarzynski M.A., et al. Glycemic variability: importance, relationship with physical activity, and the influence of exercise. Sports Med Health Sci. 2021;3(4):183–193. doi: 10.1016/j.smhs.2021.09.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kashiwagi K., Inaishi J., Kinoshita S., et al. Assessment of glycemic variability and lifestyle behaviors in healthy nondiabetic individuals according to the categories of body mass index. PLoS One. 2023;18(10) doi: 10.1371/journal.pone.0291923. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ohigashi M., Osugi K., Kusunoki Y., et al. Association of time in range with hemoglobin A1c, glycated albumin and 1,5-anhydro-d-glucitol. J Diabetes Investig. 2021;12(6):940–949. doi: 10.1111/jdi.13437. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Beck R.W., Raghinaru D., Calhoun P., et al. A comparison of continuous glucose monitoring-measured time-in-range 70-180 mg/dL versus time-in-tight-range 70-140 mg/dL. Diabetes Technol Therapeut. 2023 doi: 10.1089/dia.2023.0380. [DOI] [PubMed] [Google Scholar]
- 10.Inaishi J., Saisho Y., Watanabe Y., et al. Changes in glycemic variability, gastric emptying and vascular endothelial function after switching from twice-daily to once-weekly exenatide in patients with type 2 diabetes: a subpopulation analysis of the twin-exenatide study. BMC Endocr Disord. 2022;22(1):20. doi: 10.1186/s12902-022-00932-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Tentolouris N., Argyrakopoulou G., Katsilambros N. Perturbed autonomic nervous system function in metabolic syndrome. NeuroMolecular Med. 2008;10(3):169–178. doi: 10.1007/s12017-008-8022-5. [DOI] [PubMed] [Google Scholar]
- 12.Reutrakul S., Van Cauter E. Sleep influences on obesity, insulin resistance, and risk of type 2 diabetes. Metabolism. 2018;84:56–66. doi: 10.1016/j.metabol.2018.02.010. [DOI] [PubMed] [Google Scholar]
- 13.Figueira F.R., Umpierre D., Bock P.M., et al. Effect of exercise on glucose variability in healthy subjects: randomized crossover trial. Biol Sport. 2019;36(2):141–148. doi: 10.5114/biolsport.2019.83006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Gude F., Díaz-Vidal P., Rúa-Pérez C., et al. Glycemic variability and its association with demographics and lifestyles in a general adult population. J Diabetes Sci Technol. 2017;11(4):780–790. doi: 10.1177/1932296816682031. [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.