Sleep and mood disorders in women with dry eye disease

The aim of the present study was to evaluate sleep and mood disorders in women aged 30–69 with dry eye disease (DED). All subjects underwent corneal examinations, with 890 completing a questionnaire regarding symptoms of DED and 213 completing both the Pittsburgh Sleep Quality Index (PSQI) and the Hospital Anxiety and Depression Scale (HADS) questionnaires. Subjects were then divided into three groups based on age (younger [30–45 years], perimenopausal [46–55 years], and older [56–69 years]), and comparisons were made among groups in subjects with and without DED. PSQI scores were significantly worse in subjects with (6.1 ± 2.9) than without (4.9 ± 2.7) DED (P = 0.003) and, in the younger group, HADS scores were worse in those with (13.2 ± 6.0) than without DED (9.7 ± 6.0) (P = 0.020). In contrast, there were no differences in mood indices between those with and without DED in the other groups. PSQI score was significantly correlated with HADS rather than ocular findings. In conclusion, sleep quality had deteriorated in women with DED. However, mood problems contributed more to sleep quality than ocular status, especially in those with DED in the younger group.


Inclusion and exclusion criteria for people with eye disease and systemic comorbidities.
The present study recruited consecutive female patients, aged between 30 and 69 years with best corrected visual acuity equal to or better than 20/25 in both eyes, from participating clinics during the study period. Subjects with any corneal disease, cataract, fundus pathology, including macular diseases and diabetic retinopathy, and glaucoma, all of which may reduce visual function, were excluded from the study. In addition, patients with diagnosed psychiatric disease and taking specific psychiatric medications were excluded from the study. Statistical analysis. Subjects were divided into three groups based on age (younger [30-45 years], perimenopausal [46-55 years], and older [56-69 years]), and then into those with and without DED, served as control. Results regarding sleep and mood disorders were compared between those with and without DED in these different age groups; in Japan, 50 years is the mean of age at which menopause occurs 31 . Results of the Schirmer test and the tear BUT in the right eye were used for analysis.
Where appropriate, data are given as the mean ± SD. The HADS and PSQI scores in each group were calculated and compared among the specified groups. To identify which ophthalmic parameters were correlated with psychiatric problems (sleep and mood disorders) in DED, regression analysis was performed for those with DED, with psychiatric parameters as dependent variables and demographic (age) and ophthalmic parameters as independent variables. All analyses were performed using StatFlex (Atech, Osaka, Japan) with P < 0.05 considered significant.

Results
In all, 1506 patients provided responses for the DED symptoms interview, with 890 of these aged between 30 and 69 years; 305 were aged 30-45 years (younger group), 255 were aged 46-55 years (perimenopausal group), and 330 were aged 56-69 years (older group). Within the younger, perimenopausal, and older groups, 161 (52.8%; mean age 38.2 years), 141 (55.3%; mean age 50.8 years), and 169 (51.2%; mean age 63.4 years), respectively, had DED. The PSQI and the HADS questionnaires were completed by consecutive 900 female patients in eye clinics. Of these, 106 and 107 female patients with and without DED, respectively, aged between 30 and 69 years of age met the criteria in the present study (Table 1). DED medication is also listed in Table 1.
Analysis of symptoms and signs revealed significant differences between the younger and older groups for the presence of photophobia and blurring (P = 0.03, P = 0.02, respectively, Chi-squared test), but not for the presence of dryness, eye fatigue, pain, or irritation (Fig. 1). Dryness and eye fatigue were the most prevalent symptoms of DED. Ocular surface examinations did not reveal any significant differences among three age groups (Fig. 2).
The PSQI scores were significantly worse in the DED than non-DED group (P = 0.003, Mann-Whitney U-test), and this was most evident in the older DED group (P = 0.005) ( Fig. 3a; Table 2). In the younger group, Scientific RepoRts | 6:35276 | DOI: 10.1038/srep35276 HADS scores were worse in the DED than non-DED group (P = 0.02), but there was no significant difference between those with and without DED in the other age groups (Fig. 3b). The PSQI subscales revealed significantly longer sleep latency (P = 0.049, Mann-Whitney U-test) and worse subjective sleep quality (P = 0.01) in DED compared with non-DED subjects in the younger group (Table 2). In the older group, subjects with DED had a shorter sleep duration (P = 0.001) and later bedtime (P = 0.000) than those without DED. Analysis of mood indices revealed that HADS scores (P = 0.02), anxiety subscores (P = 0.02), and depression subscores (P = 0.02) were significantly worse in those with than without DED in the younger group. In contrast, there were no significant differences in those with and without DED in the other age groups. Younger women with DED had significantly worse scores for some items on the HADS than women without DED (Mann-Whitney U-test; Fig. 4).

Discussion
Sleep quality was significantly worse in women with than without DED. Mood disorders of depression and anxiety were major contributory factors, especially in the younger DED group. There was no indication in the results of regression analysis that ocular surface problems were directly associated with sleep quality.   There was no significant difference in non-visual symptoms among the three age groups with DED. Regarding visual symptoms, photophobia (*P = 0.02, Chi-squared test) and blurring ( † P = 0.02) were more prevalent in the older than younger group with DED. There was no significant difference in the prevalence of fatigue. Generally, depression worsens with age 18,19 ; however, the results of the present study indicated that the depression score was worst in younger women with DED and, conversely, improved with age. In contrast, in women without DED, the results were comparable with previous reports, namely that sleep and mood decline gradually with age and are worst in the climacteric period. Labbé et al. reported that subjects with DED (≥ 40 years, mean age 64.8 years) were more depressive than those without DED, and that the level of depression was correlated with ocular symptoms, but not tear function, including results for the tear break-up time (BUT) and Schirmer test 13 . The results of large nation-wide survey using SF-36, validated and widely used measure for quality of life 32 , were supportive to both our findings and previous reports that scores of mental health was worst in young women aged 20-29, better with age, best in the age of 50-69, and decline after 70. Initially we hypothesized sleep and mood was worst in peri-or post-menopausal period and DED was an exacerbating factor for them and it was exactly the case with sleep quality in older group. In addition, DED affected sleep and mood most remarkably in younger group, and was not so much in peri-menopausal group, as shown in Fig. 3. Peri-menopausal symptoms depend on physical and psychosocial (personality, family, parents, colleagues) factors 33 and it is reasonable that contributory factors to symptoms may vary at each life cycle. Further investigations are necessary to clarify the severity and interaction of DED and mood disorder in pre-menopausal women since many previous studies were for middle or older women.  PSQI scores were significantly worse in women with than without DED in both the younger and older groups (*P = 0.001 and *P = 0.005, respectively). HADS scores were worse in younger subjects with than without DED (*P = 0.02), whereas there were no significant differences among the perimenopausal and other groups. It was also significantly worse in younger group than older group with DED ( † P = 0.02). *P < 0.05 compared with subjects with DED in the same age group; † P < 0.05 compared with older subjects with DED, Mann-Whitney U-test with Bonferroni correction. The present study is the first to describe DED as a possible contributory factor to sleep disorders in women in midlife. It was evident in pre-and post-menopausal period rather than peri-menopausal. According to statistics from a coding database 12 , many disorders are often complicated with DED, including sleep apnea (odds ratio    [OR] 2.46), depression (OR 1.92), and post-traumatic stress disorder (OR 1.92), and all of these could potentially exacerbate sleep quality in subjects with DED 11 . Insomnia, depression, and DED may be inter-related with each other in peri-menopausal period 34 .
In the present study, women in the older group reported more blurring and photophobia, which may be consequences of early cataract and presbyopia. This is another reason for the high prevalence of DED in the older population in the present study, because these conditions are supposedly linked with symptomatic DED, even though the visual acuity was better than 20/25 for women in the present study. There was no significant difference in corneal findings and the results of tear function tests between the younger and older groups. Older women with DED may suffer more from visual than non-visual symptoms.
In an attempt to compare DED patients with non-DED subjects without the possibility of confounding effects of systemic comorbidities, patients with glaucoma, cataract, visual impairment, and other ocular comorbidities that could affect sleep and mood were excluded from the study. Blindness, cataract 35 , and glaucoma 36 can potentially exacerbate sleep and mood and recent studies 35,36 have used pupillary examinations, polysomnography (electroencephalogram), and actigraphy in addition to questionnaires, and have revealed considerable associations between eye diseases and psychiatric status. Glaucoma is a very common eye disease that requires lifelong topical medication, which has known ocular surface toxicity 37 , and is often complicated by DED. Glaucoma patients are highly likely to experience sleep disorders 36 , and this is why glaucoma patients were excluded from the present study.
The present study has several limitations. The non-DED group consisted of eye clinic visitors who were diagnosed as having healthy vision and no ocular surface disease and normal biomicroscopic findings, although the Schirmer test and vital staining should have been performed to completely exclude subjects with DED from this group. Thus, the findings of the present study need to be confirmed using a comprehensive DED classification, such as the one proposed by the 2007 Dry Eye Workshop 38 . In addition, sleep and mood disorders should have been diagnosed precisely by qualified psychiatrists using polysomnography and actigraphy. Another limitation of the study is that it lacks sufficient consideration of systemic comorbidities and menopause status: because of time and space restrictions, we did not obtain a detailed medical history, but comorbidities and menopause status are important contributory factors to DED and sleep and mood disorders. This may have affected the results and further investigations that take comorbidities, socioeconomic status, education, and other possible confounding factors into account are needed to obtain a detailed understanding of the causal relationship between DED and psychiatric disorders.
In conclusion, sleep quality had deteriorated in women with DED. Mood problems contributed more to sleep quality than ocular status, especially in those with DED in the younger group. The findings of the present study may provide further information regarding mental health in women with DED.