Abstract
Aims/hypothesis
The aim of this study was to explore whether diabetic retinopathy is associated with alterations of the circadian system, and to examine the role of reduced intrinsically photosensitive retinal ganglion cell (ipRGC) function.
Methods
Participants with type 2 diabetes, with diabetic retinopathy (n=14) and without diabetic retinopathy (n=9) underwent 24 h blood sampling for melatonin and cortisol under controlled laboratory conditions. ipRGC function was inferred from the post-illumination pupil response (PIPR). Habitual sleep duration, efficiency and variability were assessed by actigraphy.
Results
Participants with diabetic retinopathy compared to participants without diabetic retinopathy had smaller PIPR (p=0.007), lower 24 h serum melatonin output (p=0.042) and greater day-to-day sleep variability (p=0.012). By contrast, 24 h cortisol profiles, sleep duration and efficiency were similar in both groups. Six individuals with diabetic retinopathy had no detectable dim-light melatonin onset. PIPR correlated with 24 h mean melatonin levels (r= 0.555, p=0.007).
Conclusions/interpretation
ipRCG dysfunction in diabetic retinopathy is associated with disruptions of the 24 h melatonin rhythm, suggesting circadian dysregulation in diabetic retinopathy.
Keywords: Circadian, Diabetic retinopathy, Melatonin
Introduction
Traditionally viewed as a microvascular disease, recent studies have revealed that the retinal neurovascular unit is affected early in the course of diabetic retinopathy (1). The intrinsically photosensitive retinal ganglion cells (ipRGCs), which express the photopigment melanopsin, located in the inner retina, can be compromised in diabetic retinopathy (2, 3). ipRGCs are a crucial component of the entrainment of the human circadian system to the light–dark cycle and play a role in synchronising 24 h rhythms, including the secretion of melatonin by the pineal gland during the biological night (4, 5). Melatonin is a neurohormone that influences sleep and glucose metabolism, with its receptors detectable in pancreatic beta cells (6). Thus, ipRGC dysfunction could result in dysregulation of melatonin, glucose and circadian rhythms, affecting non-visual health in diabetic retinopathy.
Recently, we reported that reduced ipRCG function in individuals with type 2 diabetes who had diabetic retinopathy was associated with lower nocturnal secretion of urinary 6-sulfatoxymelatonin and an absence of a detectable dim-light melatonin onset (DLMO) as assessed by evening salivary sampling, providing preliminary evidence that ipRCG dysfunction may alter melatonin regulation and impact the circadian system (7). However, it was unknown if evening saliva sampling failed to capture the DLMO, which may have been mistimed in participants with diabetic retinopathy.
The aim of this study was to examine in detail the link between ipRCG function and the 24 h serum melatonin and cortisol profiles, two widely recognised markers of circadian regulation, as well as sleep duration, quality and day-to-day variability in individuals with type 2 diabetes, with and without diabetic retinopathy.
Methods
Study population
Patients age 40–65 years with type 2 diabetes being followed at the ophthalmology clinic at the University of Illinois Chicago (UIC) were recruited to participate in a study of sleep, hormone rhythms and ipRCG function. Participants with diabetic retinopathy were a part of a clinical trial and their baseline data were utilised (NCT04547439, February 2020–October 2022), while participants without diabetic retinopathy had similar data collection procedures but did not enter the clinical trial. Patients without diabetic retinopathy (n=9) and patients with non-proliferative diabetic retinopathy of at least moderate severity (n=14), based on an ophthalmological examination (per the Early Treatment of Diabetic Retinopathy Study Scale [ETDRS]) (8) within one year of enrolment, were included. Data of one diabetic retinopathy patient from a previous study were also included (7). Participants were recruited from the ophthalmology clinic at UIC. Exclusion criteria included but were not limited to pregnancy, significant medical and psychiatric comorbidities, night shift work and use of exogenous melatonin or illicit drugs, a history of ischaemic optic neuropathy, optic nerve diseases, glaucoma, retinal vascular disorders or panretinal photocoagulation. The protocol was approved by the Institutional Review Board at both UIC and the University of Chicago. All participants gave a written informed consent.
Demographic information (self-reported sex/gender, race/ethnicity; the distribution of which were representative of patients seen at UIC, Table 1), diabetes history, medication use, weight and height, were obtained. Income information was not collected. Visual acuity (log minimum angle of resolution, log MAR) was obtained from the ophthalmological examination above. HbA1c, serum eGFR, and urine albumin/creatinine ratio were collected and assayed by Quest Diagnostics (Chicago, IL, USA). Foot examination was performed by the research team to assessed peripheral neuropathy (vibratory sensation at great toe [C128 tuning fork] and 10g monofilament sensation) and recorded as normal, decreased or absent.
Table 1.
Characteristics of participants
| Characteristic | Diabetes without DR (n=9) | Diabetes with DR of at least moderate severity (n=14) | p value |
|---|---|---|---|
| Age, years | 53.6±7.5 | 52.9±6.9 | 0.840 |
| Women, n (%) | 5 (55.5) | 8 (57.1) | 0.940 |
| BMI, kg/m2 | 31.3±4.4 | 35.7±7.1 | 0.112 |
| Ethnicity, n (%) | 0.183 | ||
| Non-Hispanic | 7 (77.7) | 7 (50.0) | |
| Hispanic | 2 (22.2) | 7 (50.5) | |
| Race, n (%) | 0.383 | ||
| White | 2 (22.2) | 6 (64.2) | |
| Black | 5 (55.5) | 5 (35.7) | |
| Asian | 1 (11.1) | 0 (0.0) | |
| Other | 1 (11.1) | 3 (21.4) | |
| Diabetes duration, years | 13.9 (8.9) | 15.0 (6.6) | 0.736 |
| Serum eGFR, ml/min per 1.73m2 | 88.2±22.8 | 76.7±25.3 | 0.283 |
| Urine ACR ≥3.39 mg/mmol, n (%) | 2 (22.2) | 5 (41.6)a | 0.350 |
| Glucose-lowering medication, n (%) | |||
| Insulin | 3 (33.3) | 9 (64.2) | 0.147 |
| Sulfonylurea | 2 (22.2) | 5 (35.7) | 0.190 |
| Metformin | 7 (77.7) | 9 (64.2) | 0.493 |
| SGLT2i | 3 (33.3) | 5 (35.7) | 0.907 |
| DPP4i | 0 (0) | 1 (7.1) | 0.412 |
| GLP1RA | 5 (55.6) | 6 (42.8) | 0.552 |
| Anti-hypertensive medication, n (%) | |||
| ACEi/ARB | 7 (77.7) | 10 (71.4) | 0.735 |
| Calcium channel blocker | 1 (11.1) | 2 (14.3) | 0.825 |
| β-blocker | 2 (22.2) | 2 (14.3) | 0.624 |
| Diuretic | 1 (11.1) | 6 (42.8) | 0.106 |
| Statin use, n (%) | 8 (88.8) | 10 (71.4) | 0.322 |
| HbA1c (%) | 6.5 (6.1–7.9) | 7.2 (6.5–8.8) | 0.179 |
| HbA1c (mmol/mol) | 47.5 (43.2–62.8) | 55.2 (47.5–72.7) | |
| Decreased or absent vibratory sensation at great toe, n (%) | 6 (66.6) | 14 (100) | 0.021 |
| Decreased or absent monofilament sensation, n (%) | 0 (0) | 4 (28.6) | 0.078 |
| Relative PIPR | 0.329±0.086 | 0.158 (0.155) | 0.007 |
| Visual acuity (log MAR) | 0.07 (0.10) | 0.25 (0.28) | 0.039 |
| Subjective sleep and circadian variables | |||
| Pittsburgh Sleep Quality Index score | 5.6±3.9 | 7.3±4.3 | 0.251 |
| Epworth Sleepiness Scale | 6.4±3.6 | 7.4±4.5 | 0.589 |
| Composite Score of Morningness | 40.9±6.5 | 39.7±8.3 | 0.725 |
| Circadian variables from 24 h sampling | |||
| 24 h serum melatonin output, pmol/l per 24 h | 1033.8 (595.1–1520.8)b | 346.7 (178.9–992.6) | 0.042 |
| 24 h mean serum melatonin level, pmol/l | 42.3 (24.4–62.5)b | 14.2 (7.4–42.5) | 0.050 |
| DLMO, hh:mm | 21:56±01:22 | 21:41±02:58b | 0.818 |
| DLMO presence, n (%) | 9 (100) | 8 (57.1) | 0.022 |
| Melatonin peak time, hh:mm | 02:15±02:26b | 00:17±06:06 | 0.301 |
| Melatonin peak level, pmol/l | 142.5 (69.3–171.8)b | 31.7 (11.0–131.1) | 0.070 |
| Melatonin amplitude, pmol/l | 70.2 (33.5–82.3)b | 14.6 (3.9–63.1) | 0.070 |
| 24 h mean serum cortisol level, nmol/l | 172.7 (157.5–185.4)c | 164.8 (151.6–200.3) | 0.856 |
| Cortisol nadir time, hh:mm | 00:21±02:43c | 23:03±04:00 | 0.445 |
| Cortisol nadir level, nmol/l | 47.5 (43.0–55.7)c | 55.0 (38.8–70.8) | 1.000 |
| Cortisol peak time, hh:mm | 07:30 (05:30–08:30)c | 08:00 (07:30–11:15) | 0.110 |
| Cortisol peak level, nmol/l | 346.6±61.2c | 368.8±81.8 | 0.533 |
| Cortisol amplitude, nmol/l | 145.9±32.3c | 157.7±34.8 | 0.545 |
| Sleep variables from actigraphy | |||
| Sleep duration, min | 364.3±73.6 | 389.6±47.3 | 0.327 |
| Sleep efficiency, % | 85.1 (78.6–87.6) | 82.3 (77.3–85.4) | 0.336 |
| Day-to-day variability of sleep duration, min | 59.0±19.6 | 88.5±28.1 | 0.012 |
| Mid-sleep time, hh:mm | 03:14±01:02 | 03:50±01:22 | 0.281 |
| Rest–activity rhythm parameters | |||
| IS | 0.22±0.11 | 0.33±0.110 | 0.276 |
| IV | 0.94±0.23 | 0.88±0.20 | 0.483 |
| RA | 0.59±0.14 | 0.65±0.14 | 0.270 |
| L5-onset, h after noon | 10.8 (9.6–11.3) | 10.7 (8.6–14.0) | 0.159 |
| M10-onset, h after midnight | 7.1 (4.5–7.3) | 8.6 (6.1–10.5) | 0.072 |
| Sleep-disordered breathing assessment | |||
| Apnoea hypopnea index, events/h | 6.9 (3.9–17.6) | 14.4 (6.4–29.7) | 0.159 |
Data are expressed as mean ± SD for normally distributed variables, median (IQR) for non-normally distributed variables, and frequency (%)
ACEi, ACE inhibitor; ACR, albumin/creatinine ratio; ARB, angiotensin receptor blocker; DPP4i; dipeptidyl peptidase-4 inhibitor; DR, diabetic retinopathy; GLP1RA, glucagon like peptide-1 receptor agonist; SGLT2i, sodium–glucose cotransporter 2 inhibitor
n=12
n=8
n=7
ipRCG assessment
The pupillometry apparatus and methodology, performed at UIC, have been previously described (2). In brief, an LED-driven ganzfeld system was used to present stimuli, and pupil size was recorded with an infrared camera system. Dark-adapted pupil responses were elicited using red and blue 450 cd/m2 flashes. Responses to the red and blue flashes of light were normalised by the median baseline pupil size. The post-illumination pupil response (PIPR) was measured as the difference between the baseline and the median pupil size measured 5–7 s following stimulus offset. The PIPR elicited by the blue flash was divided by the PIPR elicited by the red flash, which defined the ‘relative PIPR’ that was used for all analyses. Of note, other definitions of the PIPR (2) yielded similar results.
Subjective sleep and circadian timing preference assessment
Self-reported sleep quality was assessed using the Pittsburgh Sleep Quality Index (9), and daytime sleepiness was assessed using the Epworth Sleepiness Scale (10), with higher scores indicating poorer sleep quality and more sleepiness. The Composite Scale of Morningness was used to assess morningness–eveningness preference (11), with higher score indicating a greater morning preference.
Ambulatory sleep assessments
Participants wore an Actiwatch spectrum wrist activity monitor (Philips Respironics, Bend, OR, USA) for 14 days and completed a sleep log daily. Sleep duration and efficiency were calculated using the Actiware 6.0 software. Mid-sleep time was calculated as the midpoint between sleep onset and offset. The standard deviation of sleep duration over the recording period was used to estimate day-to-day variability of sleep duration.
Sleep-disordered breathing was assessed using an FDA-approved portable diagnostic device, WatchPAT 300 (Itamar Medical, Israel) (12). If the participants were using continuous positive airway pressure (CPAP), the assessment was performed while using CPAP.
Circadian assessment
The 24 h sampling studies were conducted in the University of Chicago Clinical Research Center. Participants were limited to sedentary activities during waking hours and were housed in a private room with bedtimes designed to mimic the individual habitual schedule. No naps were allowed. Only dim-light exposure (approximately 75–100 lux, ceiling lights off, shades closed) was allowed during daytime and all lights were off during scheduled sleep times. Blood sampling was conducted every 30 min with procedures regarding collection and storage as previously described (13). Cortisol levels were measured by an immunochemiluminescent assay (Immulite, Los Angeles, CA, USA). Serum melatonin concentrations were measured by radioimmunoassay (Buhlmann RIA—SolidPhase, Portland, ME, USA). DLMO was defined as the timing of the first sample when melatonin levels were at or above 17.22 pmol/l for at least 2 h. To quantify the 24 h variation of melatonin and cortisol, a best-fit curve was calculated for each individual profile using a robust locally weighted nonlinear regression procedure with a window of 2 h (14). The peak and the nadir were defined as the maximum and minimum of the regression curve, respectively. The amplitude was defined as half of the difference between the peak and the nadir. 24 h melatonin output was defined as the total incremental AUC, calculated using the trapezoidal method.
Rest–activity rhythm analysis
The rest–activity rhythm from actigraphy-derived wrist activity recordings was assessed by nonparametric variables (15) using ‘nparACT’ package for R (16). The following variables were derived: intradaily variability (IV; reflecting the fragmentation of the rhythm), interdaily stability (IS; indicating the regularity of sleep patterns across days), M10-onset (start time of the most active 10 h period), L5-onset (the start time of the least active 5 h period), and relative amplitude (RA; a larger number reflects larger amplitude (15).
Statistical analysis
Data are presented as mean (SD), median (IQR) or frequency (%). Comparisons of variables between groups were performed with independent t test, Mann–Whitney U test or χ2 test where appropriate. Analyses of associations were performed using Spearman’s or Pearson correlations. Analyses were performed using SPSS version 28 (Chicago, IL, USA).
Results
Of the 186 potentially eligible participants, 48 were interested and confirmed eligible, 37 were enrolled and 23 underwent a complete assessment of 24 h blood sampling. Participant characteristics are shown in Table 1. Of the 14 participants with diabetic retinopathy, nine had moderate and five had severe diabetic retinopathy. Mean age, BMI, medication use, eGFR and HbA1c did not differ between participants with and without diabetic retinopathy, while those with diabetic retinopathy had more signs of peripheral neuropathy. Relative PIPR was significantly smaller in participants with diabetic retinopathy compared to those without (mean [SD]: 0.158 [0.155] vs 0.329 [0.086], p=0.007). Nightly sleep duration and efficiency were similar, but sleep duration variability was greater in participants with diabetic retinopathy than those without diabetic retinopathy (88.5 [28.1] vs 59.0 [19.6] min, p=0.012). Subjective sleep and circadian preference, and rest–activity rhythm parameters did not differ between groups.
Figure 1 shows mean profiles of 24 h serum melatonin and cortisol of participants in the no diabetic retinopathy and diabetic retinopathy groups. All participants without diabetic retinopathy exhibited the classical 24 h profile of melatonin with consistently low levels during the daytime and a clear nocturnal rise peaking around the middle of the night. While DLMO was present in all participants without diabetic retinopathy, it was detected only in eight of the 14 participants with diabetic retinopathy (57.1%). As reported in Table 1, 24 h melatonin output was lower in participants with diabetic retinopathy than in those without (p=0.042), while peak and amplitude levels were not significantly different (p=0.07). Diabetic retinopathy patients without DLMO had 24 h mean melatonin levels under 17.22 pmol/l (Fig. 1). By contrast, 24 h cortisol profiles (nadir and peak levels and timing, mean and amplitude) were roughly similar in both groups and followed the expected waveshape of cortisol secretion (Fig. 1, Table 1).
Fig. 1.
24 h profiles of melatonin and cortisol. (a) Mean 24 h melatonin profiles in participants with and without diabetic retinopathy. (b) Mean 24 h melatonin profiles in participants with diabetic retinopathy, and with and without DLMO. (c) Mean 24 h cortisol in participants with and without diabetic retinopathy. Vertical lines at each time point represent the SEM. DR, diabetic retinopathy
Correlations between variables with PIPR, 24 h melatonin output and presence of DLMO are shown in electronic supplementary material (ESM) Table 1. There was a highly significant correlation between PIPR and 24 h melatonin level (r= 0.555, p=0.007) and output (r= 0.549, p=0.008), ESM Table 1. Visual acuity also related to PIPR but not to melatonin profile. Furthermore, poorer subjective sleep quality, as assessed by Pittsburgh Sleep Quality Index, significantly associated with lower melatonin levels (p=0.005), ESM Table 1. Participants without DLMO had significantly smaller PIPR compared to those with detectable DLMO (0.101 [0.135] vs 0.268 [0.141], p=0.019), and they tended to have lower sleep efficiency (p=0.052), ESM Table 1. No other significant correlations were found.
Discussion
This study provides novel data on circadian regulation in individuals with and without diabetic retinopathy based on assessment of the 24 h profile of melatonin and cortisol, two robust markers of human circadian timing. The result demonstrated that diabetic retinopathy is associated with a markedly dysregulated melatonin rhythm, including low levels and absence of robust day–night circadian variation. Almost half of the participants with diabetic retinopathy had no DLMO and virtually no detectable melatonin concentrations across the 24 h period. The degree of dysregulation correlated strongly with the function of ipRGCs as estimated by the PIPR, supporting adverse effects of retinal neurodegeneration on circadian health in diabetic retinopathy. Lower melatonin also related to worse sleep quality. In contrast, cortisol profiles in participants with diabetic retinopathy compared to those without were similar. This discrepancy is likely due to different ‘zeitgebers’ or ‘time givers’ influencing melatonin and cortisol secretion. Impaired light–dark input into the central circadian clock, likely as a result of ipRCG dysfunction, directly affects melatonin secretion by the pineal gland. Cortisol rhythmicity, however, is influenced by additional factors including the timing of the rest–activity cycle, sleep–wake cycle and food intake (17). The observed differences between the melatonin and cortisol profiles indicate that there is some degree of misalignment between central and peripheral circadian clocks in diabetic retinopathy.
As circadian misalignment has been shown to be detrimental to metabolic and cardiovascular health (18), interventions to reduce misalignment could potentially offer health benefits in diabetic retinopathy. Building on the findings of low overnight secretion of urinary 6-sulfatoxymelatonin in diabetic retinopathy (7, 19), our results raise the question of whether appropriately timed melatonin supplementation could be beneficial to metabolic health in patients with diabetic retinopathy, a population currently estimated at 9.6 million in the USA (20). Although glycaemic benefits and reduction in inflammation have been reported following melatonin supplementation in various conditions (21, 22), no study has, until now, been conducted specifically in patients with diabetic retinopathy. Dosage, timing and coordination with timing of food intake will be important considerations in melatonin supplementation (23). Interestingly, in our participants, diabetic retinopathy was associated with greater day-to-day variability of sleep duration, a recently recognised risk factor for adverse health outcomes (24). The mechanism underlying irregular sleep duration is unclear as correlations with PIPR or melatonin profiles were not significant. Interventions to enhance circadian regulation, including appropriately timed dietary intake, exercise and regular sleep, may be particularly beneficial in diabetic retinopathy and should be explored in future research.
Although limited by a relatively small number of participants and lack of certain circadian measurements (e.g. wrist temperature), our results clearly demonstrate that diabetic retinopathy is associated with a dysregulated melatonin profile and ipRGC dysfunction, but with an essentially normal cortisol profile, suggestive of a degree of circadian misalignment between central and peripheral clocks. These data suggest that there may be adverse health effects of diabetic retinopathy, mediated by retinal neurodegeneration, in addition to its effects on visual health related to microvascular changes. Measures of retinal neurodegeneration, such as retinal thinning or ipRGC dysfunction, are not included in the standard clinical ETDRS scale that is used to stage diabetic retinopathy. Future work is needed to develop enhanced staging of diabetic retinopathy that incorporates neural and vascular abnormalities, and that may better predict which patients are likely to suffer from melatonin dysregulation and circadian misalignment. Given the increasing number of individuals affected by diabetic retinopathy, the benefits of circadian health optimisation in these patients should be urgently explored.
Supplementary Material
Acknowledgements
We would like thank all participants in this study, as well as the staff of the University of Chicago Clinical Research Center for assistance. Some of the data were presented as an abstract at the Endocrine Society meeting in June 2023.
Funding
This study was funded by the National Eye Institute (R01EY029782, R01EY026004, P30EY001792), NIDDK P30 DK020595, and The Center for Clinical and Translational Science (CCTS) UL1TR002003.
Abbreviations
- DLMO
Dim-light melatonin onset
- ETDRS
Early Treatment of Diabetic Retinopathy Study
- ipRGC
Intrinsically photosensitive retinal ganglion cell
- IS
Interdaily stability
- IV
Intradaily variability
- MAR
Minimum angle of resolution
- PIPR
Post-illumination pupil response
- RA
Relative amplitude
- UIC
University of Illinois Chicago
Footnotes
Authors’ relationships and activities SR received speaker fees from Eli Lily. The authors declare that there are no other relationships or activities that might bias, or be perceived to bias, their work.
Data availability
Data associated with this manuscript will be available from the corresponding author upon a reasonable request.
References
- 1.Torm MEW, Dorweiler TF, Fickweiler W, et al. Frontiers in diabetic retinal disease. J Diabetes Complications. 2022;37(2):108386. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Park JC, Chen YF, Blair NP, et al. Pupillary responses in non-proliferative diabetic retinopathy. Sci Rep. 2017;7:44987. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Feigl B, Zele AJ, Fader SM, et al. The post-illumination pupil response of melanopsin-expressing intrinsically photosensitive retinal ganglion cells in diabetes. Acta ophthalmologica. 2012;90(3):e230–4. [DOI] [PubMed] [Google Scholar]
- 4.Brzezinski A Melatonin in humans. N Engl J Med. 1997;336(3):186–95. [DOI] [PubMed] [Google Scholar]
- 5.Thapan K, Arendt J, Skene DJ. An action spectrum for melatonin suppression: evidence for a novel non-rod, non-cone photoreceptor system in humans. J Physiol. 2001;535(Pt 1):261–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Mulder H, Nagorny CL, Lyssenko V, Groop L. Melatonin receptors in pancreatic islets: good morning to a novel type 2 diabetes gene. Diabetologia. 2009;52(7):1240–9. [DOI] [PubMed] [Google Scholar]
- 7.Reutrakul S, Crowley SJ, Park JC, et al. Relationship between intrinsically photosensitive ganglion cell function and circadian regulation in diabetic retinopathy. Scientific Reports. 2020;10:1560. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Davis MD, Fisher MR, Gangnon RE, et al. Risk factors for high-risk proliferative diabetic retinopathy and severe visual loss: Early Treatment Diabetic Retinopathy Study Report #18. Investigative ophthalmology & visual science. 1998;39(2):233–52. [PubMed] [Google Scholar]
- 9.Buysse DJ, Reynolds CF III, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res. 1989;28(2):193–213. [DOI] [PubMed] [Google Scholar]
- 10.Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep. 1991;14(6):540–5. [DOI] [PubMed] [Google Scholar]
- 11.Smith CS, Reilly C, Midkiff K. Evaluation of three circadian rhythm questionnaires with suggestions for an improved measure of morningness. J Appl Psychol. 1989;74(5):728–38. [DOI] [PubMed] [Google Scholar]
- 12.Zou D, Grote L, Peker Y, Lindblad U, Hedner J. Validation a portable monitoring device for sleep apnea diagnosis in a population based cohort using synchronized home polysomnography. Sleep. 2006;29(3):367–74. [DOI] [PubMed] [Google Scholar]
- 13.Hanlon EC, Tasali E, Leproult R, et al. Sleep restriction enhances the daily rhythm of circulating levels of endocannabinoid 2-arachidonoylglycerol. Sleep. 2016;39(3):653–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Cleveland W Robust locally weighted regression and smoothing scatterplots. Journal of the American Statistical Association. 1979;74(368):829–36. [Google Scholar]
- 15.van Someren EJ, Hagebeuk EE, Lijzenga C, et al. Circadian rest-activity rhythm disturbances in Alzheimer’s disease. Biological psychiatry. 1996;40(4):259–70. [DOI] [PubMed] [Google Scholar]
- 16.Blume C, Santhi N, Schabus M. ‘nparACT’ package for R: A free software tool for the non-parametric analysis of actigraphy data. MethodsX. 2016;3:430–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Van Cauter E, Shapiro ET, Tillil H, Polonsky KS. Circadian modulation of glucose and insulin responses to meals: relationship to cortisol rhythm. The American journal of physiology. 1992;262(4 Pt 1):E467–75. [DOI] [PubMed] [Google Scholar]
- 18.Leproult R, Holmback U, Van Cauter E. Circadian misalignment augments markers of insulin resistance and inflammation, independently of sleep loss. Diabetes. 2014;63(6):1860–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Chen W, Cao H, Lu QY, et al. Urinary 6-sulfatoxymelatonin level in diabetic retinopathy patients with type 2 diabetes. Int J Clin Exp Pathol. 2014;7(7):4317–22. [PMC free article] [PubMed] [Google Scholar]
- 20.Lundeen EA, Burke-Conte Z, Rein DB, et al. Prevalence of diabetic retinopathy in the US in 2021. JAMA Ophthalmology. 2023;141(8):747–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Doosti-Irani A, Ostadmohammadi V, Mirhosseini N, et al. The effects of melatonin supplementation on glycemic control: a systematic review and meta-analysis of randomized controlled trials. Hormone and metabolic research. 2018;50(11):783–90. [DOI] [PubMed] [Google Scholar]
- 22.Raygan F, Ostadmohammadi V, Bahmani F, Reiter RJ, Asemi Z. Melatonin administration lowers biomarkers of oxidative stress and cardio-metabolic risk in type 2 diabetic patients with coronary heart disease: A randomized, double-blind, placebo-controlled trial. Clinical nutrition (Edinburgh, Scotland). 2019;38(1)191–6 [DOI] [PubMed] [Google Scholar]
- 23.Garaulet M, Qian J, Florez JC, Arendt J, Saxena R, Scheer F. Melatonin effects on glucose metabolism: time to unlock the controversy. Trends in endocrinology and metabolism: TEM. 2020;31(3):192–204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Sletten TL, Weaver MD, Foster RG, et al. The importance of sleep regularity: a consensus statement of the National Sleep Foundation sleep timing and variability panel. Sleep Health. 2023; 9(6):801–820 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data associated with this manuscript will be available from the corresponding author upon a reasonable request.