Association Between Eye Diseases and Sleep Duration: A Nationwide Cross-Sectional Study

Article information

Psychiatry Investig. 2025;22(9):1038-1047

Publication date (electronic) : 2025 August 14

doi : https://doi.org/10.30773/pi.2025.0117

Young-Chan Kim¹, Suhyung Kim¹, Tae-Won Kim¹, Seung-Chul Hong¹, Ho Jun Seo¹, Jong-Hyun Jeong¹, Hyun Kook Lim², Yoo Hyun Um^,1

¹Department of Psychiatry, St. Vincent’s Hospital, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea

²Department of Psychiatry, Yeouido St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea

Correspondence: Yoo Hyun Um, MD, PhD Department of Psychiatry, St. Vincent’s Hospital, College of Medicine, The Catholic University of Korea, 93 Jungbu-daero, Paldal-gu, Suwon 16247, Republic of Korea Tel: +82-31-249-7150, E-mail: cherubic712@naver.com

Received 2025 April 4; Revised 2025 June 4; Accepted 2025 July 5.

Abstract

INTRODUCTION

Sleep is an essential factor for human health, with the circadian rhythm being crucial for maintaining healthy sleep and overall well-being. Disruptions in the circadian rhythm can lead to various health issues [1], and are a major cause of insomnia [2]. Both intrinsic factors of the human body and external stimuli are important for maintaining the circadian rhythm [3]. Among these external stimuli, the visual system is the most important for regulating the sleep circadian rhythm [4].

Given the importance of sleep for human health, with 7 to 9 hours of sleep duration generally considered optimal [5], it is concerning that individuals with ophthalmic diseases (ODs) often struggle to maintain this optimal sleep duration. Diabetic retinopathy (DMR) patients are known to have shorter sleep times and higher rates of insomnia [6,7]. Similarly, glaucoma patients exhibit high rates of abnormal sleep duration and frequently experience insomnia [8,9]. Although only a limited number of studies have been conducted, age-related macular degeneration (AMD) patients also show higher rates of short sleep duration [10].

Particularly, ODs affecting the retina can lead to visual impairment and dysfunction of the intrinsically photosensitive retinal ganglion cells (ipRGCs). The ipRGCs in the retina play the most pivotal role in regulating sleep [11]. These cells receive blue light and transmit this information to the suprachiasmatic nucleus (SCN), which in turn produces melatonin [12], a hormone vital for synchronizing the body’s internal clock with the external environment [13,14]. Therefore, various ODs that affect the retina can damage these ipRGCs, leading to sleep disorders [15-17].

These findings suggest a potential pathophysiological connection between ODs and sleep, particularly in terms of alterations in sleep duration. Previous research has primarily concentrated on individual ODs and their specific relationships with sleep, often without making comparisons across different ophthalmic conditions. Furthermore, only a limited number of studies have directly assessed variations in sleep duration across multiple ODs. Conditions such as DMR, AMD, epiretinal membrane (EM), and glaucoma possess distinct pathophysiological mechanisms, which may uniquely affect sleep patterns. Investigating the differences in sleep duration associated with these various ODs could yield deeper insights into the role of the visual system in sleep regulation. Given the rising prevalence of age-related ophthalmic conditions, understanding how these diseases affect sleep is critical to improving healthcare outcomes in aging populations.

In this study, we conducted an epidemiologic analysis of the relationship between ODs and sleep duration by examining fundus photography findings and sleep data from a nationwide survey. We aimed to investigate whether sleep duration varies depending on the specific type of OD.

METHODS

The Korea National Health and Nutrition Examination Survey (KNHANES), a nationwide survey conducted annually by the Korean Ministry of Health and Welfare, provides accessible data for researchers. This study utilized data from KNHANES VIII covering 2019 and 2020. It was conducted in accordance with approval (2024-2002-0001) from the Institutional Review Board of the Catholic University of Korea.

Study participant

The 2019 and 2020 KNHANES surveyed 15,469 participants, with a participation rate of 74.0%. This study included participants aged 40 years and above who had undergone fundoscopy. Out of 9,296 eligible participants, 8,395 completed fundoscopy and were included in the analysis.

To perform further analysis, we identified patients diagnosed with DMR, diabetic macular edema (DME), AMD, EM, macular hole (MH), retinal vein occlusion (RVO), and glaucoma through fundoscopy. We specifically focused on patients with only one specific OD, excluding those with multiple overlapping conditions, and compared their sleep duration to those without that specific disease (Figure 1).

Figure 1.

Flow diagram of enrolled participants in the study. *patients with a single ophthalmic disease, excluding those with overlapping ophthalmic conditions. KNHANES, Korea National Health and Nutrition Examination Survey; DMR, diabetic retinopathy; DME, diabetic macular edema; AMD, age-related macular degeneration; EM, epiretinal membrane; MH, macular hole; RVO, retinal vein occlusion.

RESULTS

Sociodemographic characteristics of the study population

Table 1 shows the sociodemographic data of participants. Patients with ODs (DMR, DME, AMD, EM, MH, RVO, glaucoma) were, on average, about 9 years older, had lower incomes and education levels, and higher rates of divorce or widowhood. There were no differences in sex distribution, and BMI averaged 24 kg/m².

Table 1.

Sociodemographic characteristics of study population (weighted %)

Individuals with ODs were less likely to engage in regular aerobic exercise but showed no differences in smoking or high-risk drinking. They had higher rates of comorbid chronic metabolic conditions such as hypertension, dyslipidemia, and diabetes, as well as higher rates of depression and major cancers.

Association between sleep duration and ophthalmic disease in individuals with and without the disease

Unadjusted data showed that patients with ODs had a shorter weekend sleep duration of 6.99 hours compared to 7.40 hours for those without such diseases, a difference of about 0.4 hours. Weekday sleep duration was similar at 6.85 hours and 6.70 hours, respectively, showing no significant difference. After adjusting for variables, both groups had similar durations, approximately 6.5 hours on weekdays and 6.9 hours on weekends, with both experiencing 0.3–0.5 hours longer sleep on weekends (Table 2).

Table 2.

Comparison of sleep durations between individuals with and without ophthalmic diseases (weekday/weekend)

Association between sleep duration and specific ODs in individuals with and without each disease

An additional analysis was conducted to compare patients with a single specific OD to those without the disease (excluding individuals with more than one ophthalmic condition) (Table 3).

Table 3.

Comparison of sleep durations between individuals with and without a specific ophthalmic disease (weekday/weekend)

For patients diagnosed with DMR (unweighted n=178), unadjusted data showed a reduction of approximately 0.3 hours in weekend sleep duration. After adjusting for variables, DMR patients had about 0.5 hours less sleep on both weekdays and weekends. Adjusted data indicated that these patients had a weekday sleep duration of 6.12 hours and a weekend sleep duration of 6.38 hours, barely exceeding 6 hours.

For AMD patients (unweighted n=1,109), there was no significant difference in sleep duration between patients and non-patients, with weekday and weekend sleep durations around 6.8 and 7.2 hours, respectively. After adjusting for variables, sleep durations were approximately 6.5 hours on weekdays and 6.8 hours on weekends, still showing no significant difference.

For EM patients (unweighted n=579), unadjusted data showed that they had 0.4 hours less sleep on weekdays and 0.7 hours less on weekends compared to those without EM. After adjusting for variables, EM patients still had 0.3 hours less sleep on both weekdays (6.3 hours) and weekends (6.6 hours) compared to non-EM participants.

For glaucoma patients (unweighted n=300), there was no significant difference in sleep duration between those with and without glaucoma, with weekday and weekend sleep durations of 6.8 and 7.2 hours, respectively, in the unadjusted data. After adjusting for variables, sleep durations remained similar at approximately 6.5 hours on weekdays and 6.8 hours on weekends, showing no significant difference between the glaucoma and non-glaucoma groups.

DISCUSSION

To the best of our knowledge, this is the first study to examine differences in sleep duration across various types of ODs. Using fundus photography-based diagnoses, we conducted an epidemiologic analysis in a large population, and our findings revealed that the association between ODs and sleep duration varies significantly depending on the specific type of disease.

Demographic data indicated that individuals diagnosed with ODs through fundoscopy had higher incidences of chronic conditions, depression, and cancer histories. Previous research has well-documented the frequent co-occurrence of ophthalmic and systemic diseases, and it is known that chronic ODs are highly associated with depression due to decreased vision [19-21]. The older average age of patients with ODs in our study could also account for some of these differences.

Patients diagnosed with any of the studied ODs (DMR, DME, AMD, EM, MH, RVO, or glaucoma) comprised 26% (weighted) of the surveyed population aged 40 years and above residing in Korea. The prevalence was highest for AMD (15.1%), followed by EM (9.0%), glaucoma (5.5%), and DMR (3.7%). These findings are consistent with previous research [19,22-24].

Patients with ODs showed no significant difference in sleep duration compared to the control group. This finding suggests that not all retinal diseases uniformly affect sleep duration and highlights the importance of analyzing the relationship between sleep duration and specific ODs individually. Therefore, we performed an additional analysis focusing on subjects with a single OD to examine their sleep duration.

This approach revealed clear differences in sleep duration among patients with a specific single OD (DMR, EM, AMD, glaucoma) compared to those without that disease. These differences were consistent both before and after adjusting for variables, suggesting that the relationship between ODs and sleep duration varies by disease type.

Previous studies have linked DMR with insomnia, reporting increased insomnia rates from the Korean National Insurance Service Database [7] and a more than threefold rise in short sleep duration in a 2020 polysomnography-based study [6]. In our study, DMR patients showed a reduction in sleep duration of about 0.5 hours, and they did not achieve 7 hours of sleep on both weekdays and weekends, even after adjusting for variables. One plausible mechanism is circadian rhythm disruption due to loss or dysfunction of ipRGCs in the retina. Disruption of ipRGC function can impair circadian rhythms, affecting sleep onset, duration, and quality [25-27]. Moreover, significant loss of ipRGCs has been found to cause insomnia in late-stage retinal diseases [28]. A 2020 study measuring post-illumination pupil response (PIPR) and overnight urinary 6-sulfatoxymelatonin found that DMR patients had reduced ipRGC function and experienced circadian dysregulation and sleep disturbances [15]. Another recent study measuring 24-hour melatonin secretion in DMR patients revealed dysregulation in melatonin secretion [29]. This suggests that DMR causes circadian rhythm disruption, potentially through ipRGC dysfunction, leading to reduced sleep duration. Moreover, a 2013 study on a DMR mouse model examined the induction of clock genes, which regulate the human circadian rhythm, and found reduced induction of these genes in the DMR mouse model [30]. This indicates that DMR may cause circadian rhythm disruption not only through ipRGC dysfunction but also through changes in clock gene induction, leading to reduced sleep duration.

Beyond retinal circadian disruption, it is important to note that DMR typically develops in the context of long-standing diabetes and is often accompanied by systemic complications and chronic inflammation [31,32]. Systemic inflammation and metabolic dysregulation associated with advanced diabetes could independently contribute to sleep disturbances. Elevated inflammatory markers have been shown to correlate with reduced sleep duration, suggesting a potential role of chronic systemic inflammation in sleep disturbance [33]. Furthermore, the presence of diabetic complications has been independently associated with shorter sleep duration, highlighting the broader physiological burden of diabetes on sleep [34]. Conditions such as neuropathy, nephropathy, and vascular dysfunction may lead to discomfort, pain, and physiological stress that impair sleep quality and reduce total sleep time. Thus, patients with DMR may experience sleep reduction through multiple interacting pathways, including direct circadian impairment via retinal damage and indirect effects stemming from systemic diabetic pathology. These intertwined mechanisms may collectively explain the marked reduction in sleep observed among individuals with DMR in this study.

In this study, the sleep duration of patients with AMD showed no difference compared to individuals without AMD. Previous research on the relationship between AMD and sleep has been limited, and no definitive conclusions have been reached. A 2022 systematic review examining this relationship concluded that the available evidence was insufficient to draw clear conclusions [35]. Moreover, individual studies have reported conflicting results, with some associating AMD with longer sleep duration [36] and others with shorter sleep duration [10].

AMD primarily affects the outer layers of the retina, including the photoreceptors and retinal pigment epithelium, while generally sparing the inner retinal layers, such as the retinal ganglion cells- including ipRGCs- at least in the early stages [19]. A 2025 pilot study found no significant difference in chronotype distribution between patients with bilateral neovascular AMD and controls, suggesting that outer retinal damage alone may not substantially disrupt intrinsic circadian regulation [37]. This finding supports the possibility that relatively preserved inner retinal pathways, including ipRGCs, contribute to the lack of significant sleep duration changes observed in our AMD cohort.

Although ipRGC dysfunction has been observed in AMD, its severity appears to vary with disease progression. While one study reported subtle ipRGC impairment even in early AMD based on PIPR measurements [38], most evidence suggests that pronounced ipRGC dysfunction becomes more evident as AMD advances to later stages [17,39]. Notably, a 2023 study showed that PIPR reductions were correlated with poor sleep efficiency among late-stage AMD patients, indicating that severe retinal degeneration may impair light signaling to the SCN, leading to circadian rhythm disruption [40]. Thus, substantial ipRGC dysfunction and associated sleep disturbances are more likely to emerge in advanced AMD, whereas in early-to-intermediate stages, where inner retinal structures remain relatively intact, the impact on circadian regulation and sleep duration may be minimal [41]. This pathophysiological distinction likely explains why no significant differences in sleep duration were observed among our AMD patients, who had relatively mild disease progression.

Moreover, in early AMD, only mild structural changes such as drusen deposits typically occur in the central retina, while visual acuity tends to remain relatively well preserved [19]. Therefore, patients with early-stage AMD may not experience substantial psychological distress related to vision loss, and their psychological burden may be comparable to that associated with normal aging. However, in a study that examined ipRGC function in late-stage AMD patients, participants with advanced AMD showed significantly higher depression scores compared to controls [17], suggesting that as AMD progresses and vision loss becomes more profound, its negative impact on mental health may also worsen, potentially exacerbating sleep disturbances. In population-based studies like ours, which included AMD patients across all disease stages without stratifying by severity, it is likely that early- to intermediate-stage cases made up the majority. Consequently, the relatively mild effects in early-stage patients could have diluted the more severe impacts seen in advanced-stage patients, resulting in an overall nonsignificant association between AMD and sleep duration. These findings underscore that the impact of AMD on sleep may vary by disease stage, and that the absence of disease stratification in epidemiological analyses like ours may mask stage-specific differences.

No prior studies have explored the association between EM and sleep duration. As this is the first report of such an association, multiple mechanisms may be involved. EM is a fibrocellular tissue on the inner retinal surface, and by exerting traction, it could potentially disturb the retinal structure and indirectly affect the function of inner retinal cells, including ipRGCs [42]. This mechanical disruption may plausibly impair ipRGC function through traction force and structural distortion, contributing to circadian dysregulation. In addition, EM development is associated with local inflammation, characterized by the release of cytokines and inflammatory mediators within the retina [43].. Such localized inflammatory processes could further compromise the health and function of retinal cells, including ipRGCs, potentially exacerbating circadian rhythm disturbances.

Additionally, EM often causes noticeable visual symptoms, such as metamorphopsia, double vision, and decreased visual acuity, earlier in the disease course than DMR, AMD, or glaucoma [42]. This early onset visual discomfort may result in considerable psychological stress, potentially contributing to reduced sleep duration. Previous studies have shown that many people experience depression and anxiety due to visual impairment [44,45], and this psychological stress is a well-known cause of insomnia [46]. Although this study adjusted for the diagnosis of depression, it may not have been sufficient to fully reflect the psychological difficulties experienced by the patients. These factors could explain the reduced sleep duration observed in EM patients in this study.

The higher prevalence of sleep disorders among glaucoma patients has been attributed to ipRGC dysfunction [47]. Previous studies have confirmed reduced ipRGC function in glaucoma patients [48]. A 2015 cross-sectional study found that lower ipRGC responses (measured by PIPR) correlated with poorer polysomnography-measured sleep quality in a small glaucoma cohort [16], and a 2011 study on advanced glaucoma similarly observed reduced ipRGC function using PIPR [49]. Moreover, large-scale and clinical studies have shown that sleep duration and quality are more adversely affected in advanced glaucoma with visual field defects than in early-stage glaucoma. A 2019 big data analysis found that patients with optic disc-defined glaucoma had longer sleep duration, while those with visual field loss—reflecting more advanced disease—had shorter sleep duration [20]. Similarly, a study comparing stable and progressive primary open-angle glaucoma (POAG) found that circadian rhythm disruptions led to about 55 minutes less sleep in progressive POAG patients compared to stable ones [50]. Therefore, glaucoma not only affects retinal ganglion cells, including ipRGCs, providing a biological basis for circadian rhythm disruption, but also involves psychosocial factors such as stress related to visual impairment or reduced daily functioning, which may contribute to sleep disturbances, especially in more advanced cases. Supporting this notion, previous research has demonstrated that subjective sleep quality in glaucoma patients is significantly associated with both visual field impairment and psychological factors, including anxiety and depression [51]. In this study, glaucoma was defined solely based on optic disc findings from fundus photography as part of a nationwide epidemiologic survey. This likely reflects predominantly early-stage glaucoma, with minimal ipRGC loss and preserved visual function. Consequently, the absence of a significant association between glaucoma and sleep duration in our analysis may be due to the mild nature of disease in this population. Taken together, our findings do not contradict prior studies demonstrating sleep disruption in advanced glaucoma. Rather, they highlight that the relationship between glaucoma and sleep is stage-dependent and may not be apparent in early or subclinical disease as identified through fundus-based epidemiologic screening.

In summary, the impact of OD on sleep appears to be disease-specific and multifactorial. Retinal neurophysiology, systemic factors, the degree of vision loss, and the psychological burden associated with eye disease may all contribute to how sleep is affected. Our discussion for each condition reflects this complexity, which may explain why only certain diseases, DMR and EM, were associated with reduced sleep in this population-based sample.

This study has several limitations. First, as a cross-sectional study, it cannot establish causal relationships. The relationship between sleep and ODs is likely bidirectional. Previous research suggests that sleep disorders can exacerbate DMR [52]. Additionally, AMD and sleep dysfunction likely have a bidirectional relationship [35]. Short or long sleep durations are also identified as risk factors for glaucoma [9]. Second, our disease classifications were based solely on fundus photographs without confirmatory tests such as optical coherence tomography or visual field testing. This approach, although appropriate for epidemiologic screening, may have led to misclassification or underdiagnosis of certain conditions. Thus, our findings reflect associations with fundus-photo-detected disease rather than confirmed clinical diagnoses. Third, this study did not examine the circadian rhythms of the patients. ODs can significantly impact circadian rhythms, but this study did not address this aspect. Nonetheless, given that circadian dysregulation is a known cause of insomnia [2], examining sleep duration changes related to ODs in this study remains meaningful. Fourth, this study adjusted only for depression, which may not sufficiently account for vision-related psychological factors such as stress, anxiety, or insomnia. Moreover, sleep duration was assessed using self-reported survey data rather than objective measures like polysomnography, raising concerns about potential reporting bias. Lastly, this study did not account for visual acuity, disease severity, or laterality. Given that ODs are progressive and that bilateral involvement typically causes greater visual impairment than unilateral disease, the lack of such detailed information may have attenuated the observed associations between eye diseases and sleep duration. Subtle gradations in disease severity and extent were not captured due to dataset limitations.

In conclusion, in this nationwide population-based study using fundus photography diagnoses, we identified disease-specific associations between ODs and sleep duration. Specifically, DMR and EM were significantly associated with reduced sleep duration, while AMD and glaucoma were not.

These findings underscore the epidemiologic relevance of assessing sleep duration across different ODs in the general population. Using objective retinal imaging data, our study provides real-world evidence of differential sleep impact according to disease type. This highlights the need for disease-specific approaches when addressing sleep health in patients with ODs. Further epidemiologic and basic research is warranted to validate and expand upon these findings, potentially informing population-level screening and targeted interventions.

Supplementary Materials

Article information Continued