Receiver operating characteristic curve based on different Schirmer test thresholds. Area under the curve, 0.580; P = .001.
Han SB, Hyon JY, Woo SJ, Lee JJ, Kim TH, Kim KW. Prevalence of Dry Eye Disease in an Elderly Korean Population. Arch Ophthalmol. 2011;129(5):633–638. doi:10.1001/archophthalmol.2011.78
To evaluate the prevalence of and risk factors for dry eye disease (DED) in an elderly Korean population.
This population-based study in Yongin, Korea, included 657 individuals 65 years or older randomly selected from an official household registration database. Dry eye symptoms were assessed using a 6-item questionnaire, and DED was defined as having 1 or more dry eye symptoms often or all the time. Dry eye tests, including the Schirmer test, tear film break-up time measured using fluorescein, cornea fluorescein staining, and examination for meibomian gland dysfunction, were performed.
The crude prevalence of DED was 30.3%, and the age-, sex-, and region (urban vs rural)–adjusted prevalence was 33.2%. Women (P = .01) and urban dwellers (P = .001) were more likely to have DED. Of those with DED, 85.1% had tear film break-up time of 10 seconds or less, 54.1% had meibomian gland dysfunction, 39.2% had a fluorescein score of 1 or greater, and 32.8% had a Schirmer test score of 5 mm or less. A low Schirmer score was correlated with increased prevalence of DED, although sensitivity was low.
Dry eye disease is common in elderly Koreans; female sex and living in an urban region were strongly correlated with its frequency. We also noted a positive correlation between low Schirmer test scores and symptom-based DED.
Dry eye disease (DED) is one of the most common ocular diseases. Ocular symptoms of DED vary from mild discomfort to severe ocular fatigue and pain that may affect daily activities.1- 5 Several population-based studies2- 4,6- 15 have reported the prevalence of DED to range from 4.3% to 73.5%. The large disparity in the reported prevalence is attributed to the lack of a standardized diagnostic definition and differences in study populations.4 Although there is much inconsistency in the prevalence of DED in the literature, risk factors, such as old age, female sex, and hazardous environment, including air pollution and indoor air problems at work, are assumed to be associated with the prevalence of DED.4 The results of previous studies16- 19 have also suggested ethnic differences in the prevalence of DED between white and Asian populations, although a direct comparison of such studies involves certain limitations, such as discrepancy in the definition of DED and a difference in the age group.
Meanwhile, information is limited on the prevalence of DED in certain population groups, including those in Korea, although the clinical importance of the disease seems to be high considering the adverse effects on quality of life.1- 5 In this study, we aimed to assess the population-based prevalence of and risk factors for DED in an elderly Korean population.
This study was conducted as a part of the Yongin Aging Study, a population-based longitudinal cohort study concerning the mood, memory, and sensory functions of elderly Korean individuals 65 years or older. This study was approved by the institutional review board of the Seoul National University Bundang Hospital and was conducted after obtaining informed consent in accord with the institutional guidelines for experimental investigations. The baseline visit for the Yongin Aging Study was conducted between May 1, 2008, and February 28, 2009, in Yongin, Korea; visits every 2 years are planned for the next 4 years, and long-term follow-up is under consideration. Yongin is a city located in the southern Gyeonggi province. In 2008, the total population of Yongin was 815 960, and 7.7% of the population was 65 years or older. Yongin was considered an ideal community for a population-based study because it is one of the largest cities in Korea with a rural/urban mixture. Then, regions within Yongin were randomly selected to cover approximately 10 000 residents 65 years or older, and 10% of the population was chosen through systematic random sampling based on residential rosters. Urban regions included the communities in the downtown of the city, and rural regions included the communities located approximately 20 km from the city center, where most inhabitants' livelihood is agriculture. Consequently, 1060 elderly individuals were included in the study sample. Invitation letters were sent to them stating that a general mental and ophthalmologic health evaluation would be performed. A specific statement about evaluation for DED was not included in the letter to avoid selection bias. Of the 1060 individuals, 657 (632 [96.2%] living independently in communities and 25 [3.8%] living in nursing homes or hospitals) agreed to participate in this study and completed a symptom questionnaire administered by trained interviewers. Interviews were performed at town offices or public health centers on a designated date. For participants with ambulation difficulty, the interviewers visited their homes or nursing homes or hospitals.
This study used a 2-stage design to estimate the DED prevalence. At the initial field survey, the Dry Eye Questionnaire (DEQ), which included 6 questions pertaining to dry eye symptoms, was administered to all the survey participants. The DEQ used at the initial field visit was created by adopting and modifying several dry eye symptom questions previously reported in the literature (Table 1).2,9,17,18 When a participant indicated the presence of a symptom, he or she was asked to specify whether the symptom was experienced rarely, sometimes, often, or all the time. Dry eye disease was defined as having 1 or more symptoms often or all the time during the past 2 weeks.
The Schirmer I test score was also determined. Without the instillation of topical anesthetics, a precalibrated dry filter strip (ColorBar; EagleVision, Inc, Memphis, Tennessee) was placed temporally in each lower fornix. The strips were removed after 5 minutes, and the amount of wetting was recorded.
All 657 subjects who completed the initial field survey were asked to visit the Seoul National University Bundang Hospital and to undergo further ophthalmologic evaluations for dry eye signs. The diagnostic status of DED was masked to all the study participants. In the secondary evaluation, 21.2% of the enrolled individuals (139 of 657) (13.1% of those asked to participate in this study [139 of 1060]) visited the hospital.
Secondary ophthalmologic examinations included slitlamp examination, tear film break-up time (BUT) testing, fluorescein staining, and meibomian gland examination. All the tests were performed by an experienced cornea specialist without knowledge of the results of the initial field visit. The tear film BUT test was performed before the other tests to avoid interference from these tests. Tear film BUT was measured using fluorescein 3 times for each eye, and the average value was used. Fluorescein staining of the cornea, which indicates corneal damage, was graded as 0 (no staining), 1 (mild staining limited to less than one-third of the cornea), 2 (moderate staining of less than half of the cornea), or 3 (severe staining of half or more of the cornea).2,9
Meibomian gland obstructions and quality of meibomian gland secretions were also graded (grade 0 indicates no obstruction and clear meibum; grade 1, plugging with translucent serous secretion when compressing the lid margin; grade 2, plugging with viscous or waxy white secretion when compressing the lid margin; and grade 3, plugging with no secretion when compressing the lid margin) and recorded.2,9 Meibomian gland dysfunction (MGD) was diagnosed as the presence of gland orifice plugging (grade ≥1).20 Abnormal clinical test cutoff points were defined as a tear film BUT of 10 seconds or less, a Schirmer test score of 5 mm or less, a fluorescein score of at least 1, or the presence of MGD, as described by other researchers.9
In each participant, the eye with worse signs was used for analyses. Determination of the eye with worse signs was made based on the clinical judgment of the examiner. The estimates of prevalence were adjusted by age, sex, and region (urban vs rural) to individuals 65 years or older in Yongin. Standardized national prevalence rates for Korean elders were also estimated using the direct standardization method, in which the prevalence rates were adjusted by age, sex, and urban vs rural region to the total Korean population, as given in the 2005 national census.
The association between demographic variables (age, sex, and urban vs rural region) and dry eye symptoms or signs was evaluated using Pearson χ2 tests and multivariate analysis using logistic regression analysis. A P < .05 was considered statistically significant. We used a commercially available software program (SPSS for Windows, version 15.0; SPSS, Inc, Chicago, Illinois) for statistical analyses.
A total of 657 individuals participated in the field survey. Their mean (SD) age was 72.0 (5.9) years (age range, 65-95 years), and the male to female ratio was 48.2%:51.8% (317:340). The Schirmer I test was completed in 648 of the 657 participants. Nine subjects refused Schirmer testing because of discomfort.
The crude prevalence rate of DED, defined as having 1 or more of the dry eye symptoms often or all the time, was 30.3% (199 of 657). The age-, sex-, and region-adjusted prevalence in elders in Yongin was 33.0% (95% confidence interval [CI], 28.8%-37.3%). The standardized prevalence in the Korean elderly population based on the 2005 national census population was 33.2% (95% CI, 28.9%-37.4%). The prevalence was significantly higher in women than in men (P = .01). Participants living in urban regions were more likely to report dry eye symptoms than were those living in rural regions (P = .001). No significant difference was noted in prevalence in the stratified age groups (Table 2). We noted an abnormal Schirmer test score (≤5 mm) in 32.8% of individuals with DED (63 of 192) and in 24.8% of those without DED (113 of 456), and the difference was significant (P = .04) (Table 3).
The secondary evaluation included 139 participants, of whom 74 (53.2%) had been diagnosed as having DED in the field survey. The group that received secondary evaluation had a significantly higher DED prevalence than did the group without secondary evaluation (odds ratio, 3.58; 95% CI, 2.43-5.28; P < .001). Of the 74 individuals with DED, 63 (85.1%) had a tear film BUT of 10 seconds or less, 29 (39.2%) had a fluorescein score of at least 1, and 40 (54.1%) had MGD (19 with grade 1, 17 with grade 2, and 4 with grade 3). Of the 65 participants who had not been diagnosed as having DED, 56 (86.2%) had a tear film BUT of 10 seconds or less, 21 (32.3%) had a fluorescein score of at least 1, and 32 (49.2%) had MGD (17 with grade 1, 14 with grade 2, and 1 with grade 3). Regarding MGD severity, there was no significant difference between participants with vs without DED (linear-by-linear association χ2 test, P = .36). With the increasing severity of MGD, the tear film BUT tended to decrease, although the difference was not significant (mean [SD] BUT: 4.73 [2.90] seconds for grade 0 MGD [n = 67 (48.2%)], 4.33 [2.88] seconds for grade 1 MGD [n = 36 (25.9%)], 3.71 [2.15] seconds for grade 2 MGD [n = 31 (22.3%)], and 3.20 [0.90] seconds for grade 3 MGD [n = 5 (3.6%)]; F = 1.329; P = .27, Welch analysis of variance).
Of all the dry eye signs, only the Schirmer test score was shown to have a significant association with positive dry eye symptoms, although the test also had a low sensitivity that only 35.8% of participants had both a Schirmer test score of 5 mm or less and positive dry eye symptoms (Table 3). Receiver operating characteristic curve analysis revealed that a Schirmer test with a cutoff value of 5 mm has a predictive value, although the accuracy is limited (area under the receiver operating characteristic curve for DED, 0.580; P = .001) (Figure). In logistic regression analysis, female sex, urban region, and low Schirmer test score were found to be associated with DED (Table 4).
To our knowledge, this is the first community-based study to investigate the prevalence of DED in Korea. Korea has a fairly homogeneous population of Asian ethnicity because there has been no mass migration in history, and, thus, is considered to be an optimal country for the evaluation of DED prevalence in an Asian population.
Although many researchers have reported the DED prevalence in various countries,2,3,6- 15 a direct comparison between such studies is difficult because of differing diagnostic criteria, study populations, dry eye symptom questionnaires, and objective tests.2,4,7 The Salisbury Eye Evaluation Study, which used a validated 6-item questionnaire, demonstrated that 14.6% of participants reported 1 or more of the 6 dry eye symptoms often or all the time.18 Schaumberg et al10,12 defined DED as the presence of clinically diagnosed DED or having severe symptoms (dryness and irritation) often or constantly and reported a prevalence of 7.8% in US women and 4.3% in US men. Other studies8,14 based on a self-administered questionnaire reported a prevalence of 28.7% (the Canadian Dry Eye Epidemiology Study) and 15.3% to 16.6% (the Blue Mountains Eye study). In Asia, previous studies2,3,7,9,11,15,16,19 that used symptom questionnaires revealed a prevalence of 21.0% to 73.5%.
In the present study, we defined DED as the presence of 1 or more dry eye symptoms included in the 6-item questionnaire occurring often or all the time (Table 1); this definition was based on current trends wherein emphasis is placed on symptom assessment in the prevalence study of DED.8,21 This emphasis is because (1) there is no gold standard among dry eye tests and DEQs have been reported to have substantial repeatability,4,22 (2) dry eye tests have been shown to have poor correlation with dry eye symptoms,4,17,18,22 and (3) the treatment goal for DED is to relieve patients' symptoms.21 Recent population-based studies6- 9,16,18,21,23 have also defined DED as the presence of frequent dry eye symptoms.
In the present study, dry eye symptoms were common in participants 65 years or older, with a sex-, age-, and region-adjusted prevalence of 33.2%. The high prevalence might, in part, be because this study included only an elderly population. Previous studies6,10,12,17 have demonstrated that the DED prevalence increases with age, although some studies8,18 have found no such correlation. However, in the present study, no significant difference was noted in stratified age groups, presumably because this study included only participants 65 years or older with a limited age distribution.
As summarized in Table 5, the DED prevalence in Asian populations is higher than that in white populations.10 A study10 conducted in the United States reported that Asian women were more likely to have severe DED symptoms than were white women, indicating ethnic differences in the prevalence of dry eye symptoms. The high DED prevalence of 33.2% in the present study is similar to the prevalence of 33.7% in elderly Chinese (≥65 years old) in Taiwan9 and considerably higher than that of 14.4% in elderly Americans.18 Uchino et al15 reported that the prevalence of DED in elderly Japanese (>60 years old) was 73.5%, suggesting that DED is highly prevalent in northeast Asia.
In the present study, female sex was associated with an increased prevalence of DED. Many previous studies have reported similar sex differences in DED prevalence,2,3,6,9,17 suggesting that sex hormones affect ocular surface conditions through their effects on lacrimal glands, meibomian glands, conjunctival goblet cell density, and ocular surface sensitivity.2,9 Individuals in urban regions were also more likely to have DED than were those in rural regions, which agrees with the findings of a recent study by Jie et al.3 Although the exact cause of this difference is unclear, the factors that contribute to the difference are assumed to include air pollution, workplace environment, and lifestyle in urban regions.3 Further studies are needed to elucidate the exact mechanism.
The lack of association between signs and symptoms of DED in the present study corresponds well with previous studies.4,17,18,22 It is conceivably due to (1) inconsistency and lack of repeatability of dry eye tests in common use, (2) the heterogenic nature of DED and variable disease process, and (3) differences in the sensation of a dry eye symptom and cognitive response to the DEQ.4 Among the dry eye tests performed, only a low Schirmer test score was shown to be associated with dry eye symptoms, and a cutoff value of 5 mm had a predictive value, although limited, for DED (Table 3 and Figure). However, the present results also show that the Schirmer test has low sensitivity, as only 35.8% of participants (63 of 176) with a Schirmer test score of 5 mm or less had positive dry eye symptoms (Table 3). The low sensitivity was also demonstrated in previous studies,17,21 in which the use of topical anesthetics was suggested to be the cause. However, anesthetics were never used in the present study, suggesting the presence of other causes, including the aforementioned factors that contributed to the discrepancy between dry eye signs and symptoms. Therefore, the Schirmer test may not be appropriate as a screening test for DED because it might be useful for aqueous-deficient dry eye but might not be predictive of a symptom-based dry eye definition. These findings agree well with those of a previous study concerning the Chinese population in Taiwan.21 A low Schirmer test score is significantly associated with dry eye symptoms in elderly Chinese and Korean populations,21 whereas no such correlation has been found in elderly white populations.17,18 This finding, along with the differences in prevalence, suggests differences in the distribution of causes of DED and susceptibility to the disease between Asian and white populations.21
Meanwhile, the results that approximately 50% of the participants had MGD suggests that MGD may be prevalent in Korea, which is in accord with the reports of other Asian countries.2,15,21 Although an association between MGD and dry eye symptoms was not found, the results suggest that MGD may cause tear film instability and worsen the ocular surface disease, and thus more attention should be given to the disease.
The present study has a few limitations. First, the severity of DED was never considered in the analyses. Investigation of more specifically graded variables—that is, dividing the “often or all the time” of the DEQ into “often” and “constantly” could yield more detailed data on the association of dry eye signs and symptoms. Second, the few patients and the higher DED prevalence in the group that underwent secondary evaluation reduced the statistical power of logistic regression analysis and, thus, could limit generalization of the study results. Although measures to avoid selection bias were used, patients with dry eye symptoms might have more incentive to visit Seoul National University Bundang Hospital and received secondary evaluation. However, that did not affect the evaluation of DED prevalence because it was determined at the initial field visit. Third, the DEQ may also be reflective of other ocular surface diseases, including MGD, allergic conjunctivitis, and chronic infectious conjunctivitis, and, thus, are not specific for DED and carry a risk of overestimating the DED prevalence. Fourth, some dry eye symptoms might not be covered by the DEQ. For example, tear film instability symptoms, including improvement in vision with blinking, were not included in the questionnaire, which possibly had an effect on the result that tear film BUT and MGD were not associated with DED. Therefore, development of the diagnostic criteria for DED that reflects the dry eye symptoms more specifically is required.
In conclusion, to our knowledge, this is the first study reporting the prevalence of DED in Korea. Dry eye disease is a common condition, with an estimated prevalence of 33.2% in elderly Koreans. Women and urban dwellers are more likely to experience dry eye symptoms. A low Schirmer test score was associated with frequent dry eye symptoms, although its diagnostic value as a single test was limited. Given the high prevalence and effect on quality of life, further studies are required to better understand the disease.
Correspondence: Ki Woong Kim, MD, PhD, Department of Neuropsychiatry, Seoul National University Bundang Hospital, 166 Gumiro, Bundanggu, Seongnamsi, Gyeonggido 463-707, Korea (firstname.lastname@example.org).
Submitted for Publication: June 4, 2010; final revision received September 17, 2010; accepted September 28, 2010.
Author Contributions: Drs Han and Hyon contributed equally to this article. All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Financial Disclosure: None reported.
Funding/Support: This study was supported by research grant 07-2008-0270 from the Ministry for Health, Welfare, and Family Affairs, Republic of Korea, and by grant A092077 from the Korean Health Technology Research and Development Project, Ministry for Health, Welfare, and Family Affairs, Republic of Korea.
Additional Contributions: Eun Ryung Han, MD, assisted with acquisition of data and clinical examination.