Sleep Disorders, Mental Health, and Dry Eye Disease: A Population-Based Study in South Korea

Dry eye disease (DED) is a multifactorial disease of the ocular surface that causes severe discomfort, mild ocular irritation, fatigue, pain, visual disturbance, and a foreign body sensation. Stress, depression, and sleep disorders are considered risk factors for DED. Our aim was to investigate the association between mental health, sleep disorders, and DED in Korean adults. This population-based, cross-sectional study examined ophthalmologic data of 16,471 Koreans aged ≥ 20 years from the 2010–2012 Korea National Health and Nutrition Examination Survey (KNHANES). DED was based on a medical diagnosis by a doctor, and data on mental health and sleep disorders were obtained using questionnaires. Multiple logistic regression analysis was conducted to examine the association between mental health, sleep disorders, and DED, and we also adjusted for possible covariates. Short sleep duration ( ≤ 5 h/night), stress perception (yes), and depressive symptoms (yes) were signicantly associated with the risk of DED even after correcting for demographic factors, lifestyle factors, and medical factors. Combined short sleep duration, stress, and depression were most strongly associated with DED, and thus ophthalmologists must be aware of the possibility of a higher prevalence of sleep disorders, stress, and depression in patients with DED.


Introduction
Dry eye disease (DED) is a growing public health problem within ophthalmology 1 . It is de ned as a multifactorial disease of the tear ducts and ocular surface, which is accompanied by increased osmolality of the tear lm and in ammation of the ocular surface 2 . The prevalence of dry eye disease among middle-aged and elderly persons ranges from 7-34% 3 . The incidence of DED, which has always been one of the most common ophthalmic diseases, has been markedly increasing in industrialized countries 4 .
DED results in severe discomfort, mild ocular irritation, fatigue, pain, visual disturbance, a foreign body sensation, and tear lm instability, which can potentially damage the ocular surface 2,5,6 . It interferes with activities of daily living 7 , thus negatively affecting patients' mental health as well as productivity at work 8, 9 . DED is recognized as a major public health problem worldwide 6 , and it can place a signi cant nancial burden on the healthcare system 10,11 .
Risk factors for DED have been reported in several studies. These risk factors include old age, female sex, smoking status, wearing contact lenses, systemic medications, video display use, and a history of ocular surgery 3,7,12−15 . Depression, sleep disorders, and stress are also considered to be risk factors for DED based on recent reports 9,16,17 . There has recently been increased medical interest in the psychiatric pathophysiology of patients with DED 18 . According to a previous study, respondents with shorter sleep durations had more DED symptoms than those with normal sleep durations 16 . However, this report has not controlled for history of ocular surgery, including refractive and cataract surgery, although these are Page 3/16 common procedures in ophthalmic clinics. DED is one of the most common complications after LASIK and cataract surgery 19,20 . Therefore, this study aimed to determine the relationship between DED and psychiatric factors of a representative sample of Korean adults aged ≥ 20 years in a population-based study using the 2010-2012 Korea National Health and Nutrition Examination Survey (KNHANES) data. In addition, we attempted to ascertain whether those associations would be affected by a comorbidity of psychiatric factors, including short sleep durations, stress perception, and depressive symptoms after adjusting for age, sex, lifestyle, and medical factors.

Study design and population
The Korean National Health and Nutrition Examination Survey (KNHANES) is a nationwide population-

Participants and data selection
We included 25,534 participants who had completed ophthalmological examinations in the KNHANES 2010-2012. We excluded participants aged <20 years (n=6,140), or those who had no DED data (n=2,599). We further excluded those without mental health data (n=324). Finally, after implementing these exclusion criteria, 16,471 participants were included for analysis in the present study ( Figure 1). DED was de ned as dry eye diagnosed by a physician, who referred to a dry eye questionnaire used in previous studies 12,9 . A questionnaire related to DED was distributed, which includes the following question: "Have you ever been diagnosed with DED by a physician?" (Participants were asked the aforementioned question with an emphasis on "by a physician."), and this question could be answered as "yes" or "no." Sleep duration was assessed via the self-administered question, "How many hours do you usually sleep a day?" The third edition of the International Classi cation of Sleep Disorders generally classi es long sleepers as those sleeping for ≥9-10 h, and short sleepers as those sleeping for ≤5-6 h 36,37 .
Therefore, this study classi ed sleep time as ≤5, 6, 7, 8, or ≥9 h/day according to the intervals used in the previous study 38 . Participants were asked to respond "yes" or "no" to the following mental health questions: "Have you ever had a lot of stress?" and "Have you ever felt depressed for 2 continuous weeks?" We de ned the participants as having a severe mental health problem when they answered positively regarding shorter sleep duration, a lot of stress perception, and depression symptomology.
Questionnaires were used to collect data on age, sex, education, household income, occupation, smoking, regular exercise, daily sun exposure, pregnancy, and history of ocular surgery. Subjects were classi ed into six age groups: 20-29, 30-39, 40-49, 50-59, 60-69, and ≥70 years. Educational attainment was classi ed into the following categories: less than an elementary school education, middle school, high school, and more than university. Household income was collapsed into quartiles. Occupation was classi ed as blue collar (agriculture workers, forestry workers, shery workers, craft and related trade workers, plant and machine operators and assemblers, and simple labor), white collar (managers, professionals, clerks, and service or sales workers), and those who were un-employed for any reason. Current smoking was classi ed as yes or no (current nonsmokers were either never smokers or past smokers). Regular exercise was de ned as walking >30 minute least 5 days a week. Daily sun exposure was classi ed into categories of <2 h, 2~5 h, and ≥5 h of typical exposure per day. Pregnancy was assessed via the question, "Are you currently pregnant?" History of previous ocular surgery was evaluated via the question, "Have you ever had ocular surgery in the past?" Anthropometric measurements were used to collect data on body mass index (BMI), diabetes, hypertension, thyroid disease, and rheumatoid arthritis. BMI was de ned as underweight BMI (<18.5 kg/m2), normal BMI (≤18.5 to <25 kg/m2) and overweight/obesity (≥25 kg/m2) 38 . Diabetes was evaluated through the question: "Have you ever diagnosed with diabetes by a physician?" Hypertension was de ned as systolic blood pressure >160 mmHg and/or diastolic blood pressure >90 mmHg, measured at the medical examination, or currently taking antihypertensive medications 38 . Thyroid disease and rheumatoid arthritis diagnoses were classi ed into combined yes or no categories through the question: "Have you ever been diagnosed with thyroid disease or rheumatoid arthritis by a physician?"

Statistical analyses
Weighted analysis was used to more accurately re ect the true population statistics. The data are expressed as unweighted numbers and weighted proportions (%). The chi-square test was used to compare the demographic, lifestyle, and medical characteristics of the study population between participants with DED and those without DED. Multiple logistic regression analysis was conducted to examine the odds ratios (ORs) and 95% con dence intervals (CIs) for the association between mental health related characteristics (sleep duration, stress perception, depressive symptoms, and severe mental health) and DED. Demographic factors (age and sex), lifestyle factors (education, occupation, BMI, smoking, regular exercise, and daily sun exposure) and medical factors (pregnancy, thyroid disease, and history of ocular surgery) were used as covariates for calculating adjusted odds ratios (aORs). Statistical analyses were performed using SPSS 18.0 version (SPSS Inc., Chicago, IL, USA). All reported P values are two-tailed, and values <0.05 was considered the threshold for statistical signi cance.

Results
Among the 25,534 people who participated in KNHANES V (2010-2012), there were 19,394 participants who were aged ≥20 years, of whom 2,599 participants did not have data on DED diagnosis and 324 participants did not have data on mental health. Those participants were excluded, leaving a total of 16,471 participants selected as nal participants (Figure 1).
In adults aged ≥20 years, the estimated DED prevalence (standard error The association between stress perception and weighted DED prevalence is presented in Figure 3. Weighted DED prevalence (SE) was 20.05% (0.92) among those with higher stress perception and 14.72% (0.57) among those with no/low stress perception. The association between depressive symptoms and weighted DED prevalence is presented in Figure 4. The association between severe mental health [combined short sleep duration (≤5), stress, and depression] and weighted DED prevalence is presented in Figure 5.
Demographic, lifestyle, and medical characteristics of participants in the present study are shown in Table 1. Regarding demographic characteristics, we found that compared with those without DED, there were many cases with DED among participants aged <20 years or those aged >50 years, among women, among those whose education level was less than an elementary school educational attainment, and among college students, unemployed participants, and white collar workers. Regarding lifestyle characteristics, there were more cases with DED among those with normal or underweight BMI, current nonsmokers, those not exercising regularly, and participants with daily sun exposure of <2 h per day compared with those without DED. Regarding clinical characteristics, there were more cases of DED among pregnant participants, those with thyroid disease, and those with previous ocular surgery.
Mental health related characteristics of subjects with and without DED are shown in Table 2. Compared with participants without DED, those with DED frequently reported a sleep duration of ≤5 h/night or ≤6 h/night in many cases, as well as a greater frequency of greater stress perception, depressive symptomology, and severe mental health symptomology. Table 3 shows the results of multiple logistic regression analyses presenting the association of DED with related mental health characteristics. Even after adjusting for demographic factors, lifestyle factors, and medical factors, people who sleep for <5 h a night (OR=1.42, 95% CI=1.06-1.90), those with greater stress perception (OR=1.71, 95% CI=1.37-2.14), those reporting depressive symptomology (OR=1.37, 95% CI=1.06-1.77), and those with severe current mental health presentation (OR=1.91, 95% CI=1.07-3.39), had a higher risk of developing DED.

Discussion
Upon exploring the correlation between DED and psychiatric factors among Koreans aged ≥ 20 years, this study showed that the risk of DED increased with short sleep duration, stress perception, and depressive symptoms. This result was in keeping with the results of previous studies. This study also showed that the risk of DED was the highest in those with severe mental health issues (those experiencing all three aforementioned factors).
DED is a common irritating eye disease that causes pain, discomfort, burning, a foreign body sensation, and blurred vision. Previous studies have evaluated the association between psychological problems and eye disease, such as age-related macular degeneration (AMD), cataract, glaucoma, and DED 21 . DED has a known association with mental health problems, such as depression and anxiety 22 . A meta-analysis evaluating the correlation between eye disease and depression showed that the prevalence of depression was 29% among patients with DED, 25% in patients with glaucoma, 24% in patients with AMD, and 23% in patients with cataract, showing the highest prevalence among patients with DED 21 .
Wan et al. 22 stated that the depression and anxiety score in patients with dry eye symptoms were higher than that in the control group, which was similar to the results obtained from this study. Ayaki et al. 23 evaluated the correlation between sleep, anxiety, and depression in 301 patients with DED and 202 patients with other ocular surface diseases (chronic conjunctivitis/ allergic conjunctivitis). They showed that the mild to severe DED group showed a higher proportion of patients presenting with problems with sleep quality based on a Pittsburgh Sleep Quality Index of ≥ 6, as well as depression symptomology based on a Hospital Anxiety and Depression Scale Depression Subscale score of ≥ 5, when compared with the chronic conjunctivitis or allergic conjunctivitis groups. Lee et al. 16 evaluated the association between sleep duration and dry eye syndrome (DES) in Korean adults. They showed that the prevalence of DES was 1.20 and 1.29-fold higher in the short sleep duration group (5 h/day) and in the severe sleep disruption group (≤ 4 h/day), respectively, compared with the optimal sleep group (6 ~ 8 h/day), showing that the prevalence of DES increased with shorter sleep duration. This is thought to be due to the stress from sensory discomfort or optical disturbance experienced by patients with DED as well as sleep di culty from in ammatory processes (pain, brightness due to in-complete eye closure, and others) during sleep. Previous studies also show that persistent positive airway pressure, a treatment option for obstructive sleep apnea, may exacerbate DED 24 .
Depression is more closely associated with dry eye symptoms compared with dry eye signs 25,26 . Irritating ocular surface symptoms exert negative effect on visual performance and daily activity 27  Galor et al. 29 stated that patients with depression and anxiety can have central sensitization, which may affect pain perception. In other words, patients with depression or anxiety may react more sensitively to ocular sensation compared to control participants 22 and experience persistent irritation from eye dryness, and this in turn can cause or exacerbate mood disorders 30 .
Furthermore, in ammatory cytokines play an important role in the pathogenesis of depression by triggering unhealthy behavior. Lymphocytes of a patient with depression activate the production of IL-1b, IL-6 and TNF-α and such in ammatory pathways are known to cause neuronal damage through oxidative and nitrosative stress 31,32 . Mrugacz et al. 33 explored the correlation between in ammatory cytokine levels in tears and depression among 32 (14 male, 18 female) DED patients with an average age of 44.21 years, observing that patients with depression have signi cantly higher in ammatory cytokines (IL-6, IL-17, and TNF-α) levels in tears than control participants, which re ects the severity of local immunological changes. In particular, locally increased IL-17 and TNF-a levels in tear uid can be considered a result of DED progression, which implies that IL-17 and TNF-a levels can be used as predictive factors for clinical in ammation in dry eye patients with depression 33 .
In some patients with DED, tear production may have been compromised from using somnifacients containing anticholinergic agents 34 . Human conjunctival epithelium contains muscarinic and adrenergic receptors, which implies that antidepressants could plausibly affect the ocular surface 38 . Chronic exposure to histamines and 5-hydroxytryptamine changes function through secretory processes, and neuronal release of 5-hydroxytryptamine could plausibly affect acute control of lacrimal secretion. Zheng et al. 21 stated that chronic painful symptoms from dry eye disease can induce depression and that medications used to treat depression can cause or exacerbate dry eye disease. Ayaki et al. 23 stated that steroid or mucin secretagogue eye drops (diquafosol and re-bamipide) used to relieve severe DED symptoms can cause distress. Sleep deprivation and medication history due to psychiatric disorders may partially inhibit tear secretion, since lacrimal secretion is under neural regulation.
Thus far, not much is known regarding the pathogenesis of DED, and it is yet unclear whether depression exacerbates DED. Ayaki et al. 18 hypothesized that DED causes distress, which can in turn give rise to depression and sleep disorders. Furthermore, eye pain and exposure to light from DED can cause sleep disturbance, which would in turn lead to sleep disorders. The inverse may also be possible, such that the presence of sleep disorder can cause DED and depression.
Irritating ocular symptoms can exert negative in uence on visual performance 23 through tear instability from DED that leads to blurred vision, which may hinder daily activities. Furthermore, DED patients can experience stress from sensory discomfort or optical disturbance, and the condition itself can cause psychological distress. DED patients experience pain and brightness due to incomplete eye closure, which can cause sleep disturbance. Such disturbance can produce in ammatory processes. Patients with Sjogren's syndrome are known to experience various sleep disorders such as pain, night awakenings, long sleep latency, and obstructive sleep apnea 18 .
A major strength of this study is the fact that it is the rst large-scale population-based study that corrects for history of ocular surgery, including refractive and cataract surgery, and evaluates the association between DED and mental health and sleep disorders. Previous studies among Asians 9,16 did not consider history of ocular surgery, although DED is the most common complication of LASIK and cataract surgery. This study was conducted among 16,471 South Korean adults aged ≥ 20 years. Korea is a single-race country with relatively uniform genetic and environmental in uences. Therefore, the results of this study are more consistent than other large population-based studies and should be broadly generalizable.
However, the following limitations must be considered. First, a cross-sectional study can identify correlation between DED and psychiatric factors, although it cannot establish a causal relationship. Second, this survey was not designed to diagnose DED, examinations related to the severity of DED were not performed. Nevertheless, many previous studies have used this data to investigate DED and its risk factors 9,12,16 . Third, the evaluation of sleep duration was performed through surveys. Therefore, future studies would require objective evaluation of sleep quality (actigraphy, melatonin measurement, and polysomnography) in patients with DED.
In conclusion, this study provides epidemiologic evidence that DED is associated with sleep duration, stress, and depression in the Korean population. Short sleep duration (≤ 5 h/night), stress perception, and depressive symptoms were signi cantly associated with the risk of DED even after correcting for demographic factors, lifestyle factors, and medical factors. When all three factors were present, in other words, in the case of severe mental health issues, the risk of DED increased greatly. Therefore, ophthalmologists must be aware of the possibility of stress and depression among patients with DED.

Declarations Data Availability
All the data supporting the conclusions of this article is included in the present article.