eMethods. Tear function test, ocular surface evaluation and sample collection
eTable1. 12 Questions of subjective symptoms for dry eye diagnosis
eTable 2. New criteria for dry eye diagnosis in Japan
eTable 3. Visual display terminal working hours and eye strain
Uchino Y, Uchino M, Yokoi N, Dogru M, Kawashima M, Okada N, Inaba T, Tamaki S, Komuro A, Sonomura Y, Kato H, Argüeso P, Kinoshita S, Tsubota K. Alteration of Tear Mucin 5AC in Office Workers Using Visual Display TerminalsThe Osaka Study. JAMA Ophthalmol. 2014;132(8):985-992. doi:10.1001/jamaophthalmol.2014.1008
There are limited reports on the relationship between mucin 5AC (MUC5AC) concentrations in tears, working hours, and the frequency of ocular symptoms in visual display terminal (VDT) users. This investigation evaluated these relationships among patients with dry eye disease (DED) and individuals serving as controls.
To determine the relationship between MUC5AC concentration in the tears of VDT users based on the diagnosis of DED and frequency of ocular symptoms.
Design, Setting, and Participants
An institutional, cross-sectional study was conducted. Participants included 96 young and middle-aged Japanese office workers. Both eyes of 96 volunteers (60 men and 36 women) were studied. Participants working in a company that used VDTs completed questionnaires about their working hours and the frequency of ocular symptoms. Dry eye disease was diagnosed as definite or probable, or it was not present. Tear fluid was collected from the inferior fornix after instillation of 50 μL of sterilized saline. The MUC5AC concentration was normalized to tear protein content and expressed as MUC5AC (nanograms) per tear protein (milligrams). The differences in MUC5AC concentration between DED groups, between VDT working hours (short, intermediate, and long), and between symptomatic and asymptomatic groups were evaluated with 95% CIs based on nonparametric Hodges-Lehmann determination.
Main Outcomes and Measures
Ocular surface evaluation, prevalence of DED, and MUC5AC concentration.
The prevalence of definite and probable DED was 9% (n = 9) and 57% (n = 55), respectively. The mean MUC5AC concentration was lower in the tears of VDT users with definite DED than in those with no DED (P = .02; Hodges-Lehmann estimator, −2.17; 95% CI, −4.67 to −0.30). The mean MUC5AC concentration in tears was lower in the group that worked longer hours than in the group that worked shorter hours (P = .049; estimated difference, −1.65; 95% CI, −3.12 to 0.00). Furthermore, MUC5AC concentration was lower in participants with symptomatic eye strain than in asymptomatic individuals (P = .001; estimated difference, −1.71; 95% CI, −2.86 to −0.63).
Conclusions and Relevance
The data obtained in the present study suggest that office workers with prolonged VDT use, as well as those with an increased frequency of eye strain, have a low MUC5AC concentration in their tears. Furthermore, MUC5AC concentration in the tears of patients with DED may be lower than that in individuals without DED.
Recent years have seen widespread use of personal computers and portable information terminals and a dramatic increase in the amount of work done using visual display terminals (VDTs). The number of people using VDTs increases globally with growing use of the Internet. In 2012, the number of Internet users worldwide was 2.4 billion, almost twice the value reported 5 years earlier.1 In Japan, approximately 95 million people (80% of the Japanese population) were using the Internet by the end of 2010.2 Prolonged VDT work has been reported to cause general and musculoskeletal symptoms, such as physical fatigue and back pain, as well as pain in the shoulders and wrists.3,4 Furthermore, VDT work has been reported5- 7 to cause ocular symptoms including asthenopia, reduced visual acuity, ocular pain, and dry eyes resulting from increased tear evaporation, decreased rate of blinking, and reduced tear secretion. In other epidemiologic studies8,9 using a questionnaire, longer duration of VDT use was associated with a significant trend toward a higher prevalence of dry eye symptoms and clinically diagnosed dry eye disease (DED). The pathogenesis of DED symptoms in VDT users may be due to changes in the composition of the tear film; however, it is unclear.
Basic research10 data suggest that the aqueous layer of the tear film is composed of mucins dissolved in the lacrimal fluid and that there is not a distinct mucin layer within the tears. The mucin in tear mucin 5AC (MUC5AC) is secreted by goblet cells in the conjunctiva.11 A large gel-forming mucin with extensive O-glycosylation, MUC5AC plays an important role in epithelial surface protection by clearing the ocular surface of debris.12 The decrease in MUC5AC concentration in ocular surface abnormalities may impair the wettability of the epithelium because the hydrophilic nature of the secreted mucins, which results from their heavy glycosylation, helps to hold fluids on epithelial surfaces.13 Data obtained via enzyme-linked immunosorbent assay have shown that the MUC5AC concentration is significantly lower in the tear fluid of patients with Sjögren syndrome than in healthy individuals.14
Although it has been reported8,9 that prolonged use of VDTs is a risk factor for DED and that MUC5AC is significantly decreased in the tears of patients with DED, there are no reports evaluating alterations of MUC5AC in the tears of VDTs users. The aim of the present combined public health– and clinical-based study was to determine the relationships between tear MUC5AC concentration in VDT users and (1) the severity of DED, (2) the number of VDT working hours, and (3) the frequency of ocular symptoms.
This research followed the tenets of the Declaration of Helsinki, and the protocol of this institutional and cross-sectional study was approved by the institutional review board of Ryogoku Eye Clinic, Tokyo, Japan. Written informed consent was received from all participants before study initiation. The participants did not receive financial compensation.
Under the supervision of the Japanese Dry Eye Society, we arbitrarily selected 2 large companies listed in the Japanese stock market and sent a letter to the industrial physicians in the health management section of each company explaining the purpose of the study and requesting their participation. The first company in the pharmaceutical sector that responded to our letter and consented to participate was enrolled in this study. The participant company was Osaka based; hence, the present study was named The Osaka Study. Participants included 96 individuals of a total of 561 who volunteered to provide impression cytology samples for The Osaka Study.
We administered a questionnaire regarding DED that is widely used in Japan.15 In brief, the questionnaire included 12 questions pertaining to the symptoms of DED. Possible answers to each question included always, often, sometimes, or never (Supplement [eTable 1]). The participants were divided into 2 groups on the basis of their answer to each of the 12 questions: those who answered always or often were placed in the symptomatic group for each ocular symptom, and those who answered sometimes or never were placed in the asymptomatic group for each ocular symptom. Information on age, sex, smoking status, height, and weight was also obtained. In addition, we categorized contact lenses use and the duration of VDT use. The participants were divided into 3 groups based on the duration of VDT use: short (<5 hours); intermediate, (5-7 hours), and long (>7 hours).
Dry eye disease was diagnosed according to the most recent diagnostic criteria of DED in Japan, described as follows (Supplement [eTable 2] ): (1) presence of DED symptoms (>1 of the 12 questions answered as always or often was considered to be indicative of positive subjective symptoms), (2) presence of qualitative and/or quantitative disturbance of the tear film (Schirmer test I value ≤5 mm and/or tear break-up time [TBUT] ≤5 seconds), and (3) presence of keratoconjunctival epithelial damage (fluorescein staining score ≥3 points and/or lissamine green staining score ≥3 points). The presence of all 3 criteria (subjective symptoms, tear fluid abnormalities, and keratoconjunctival disorder) was necessary to establish a diagnosis of definite DED. Participants positive for 2 criteria received a diagnosis of probable DED, whereas those positive for 1 criterion or negative for all 3 criteria were categorized as not having DED.9,16
All participants were asked to not use a VDT for 2 hours before the clinical examinations and sample collections, which are detailed in the Supplement (eMethods). Tests aimed at evaluating tear and ocular surface function were carried out between 9:30 am and 4:00 pm and included assessment of TBUT, Schirmer test I value, and double vital staining of the conjunctiva and cornea with lissamine green and fluorescein.9,17 Tear and impression cytology samples were collected at least 2 days following the diagnostic tests between 10:30 am and 2:30 pm, as previously described.14,18- 21
The concentration of the secreted mucin MUC5AC in the tear samples was quantified by enzyme-linked immunoassay (E90756Hu; USCN Life Science).22 All samples were analyzed according to the manufacturer’s guidelines. Absorbance was measured at 450 nm, and the standard solutions in the kit were recombinant human MUC5AC. A protein assay reagent kit (BCA Protein Assay Kit; Pierce) was used to determine the protein concentration in the tear samples. The MUC5AC concentration was normalized to the tear protein content and expressed as MUC5AC protein (nanograms) per tear total protein (milligrams).
Total RNA was isolated from the impression cytology samples using an extraction kit (RNeasy Mini Kit; Qiagen). The homogenization step was skipped because of the small number of cells, and all RNA extraction procedures followed the manufacturer’s instructions with an additional DNase I digestion step.21,23 Total RNA was used for complementary DNA synthesis (ReverTra Ace-α-; Toyobo) as described in the manufacturer’s protocol. Reverse transcription was performed by incubating the tubes at 30°C for 10 minutes, 42°C for 20 minutes, 99°C for 5 minutes, and 4°C for 5 minutes. The reverse transcription products were then ready for use in real-time polymerase chain reaction (PCR).24 A total of 2 µL of the total reverse transcription volume was used for quantitative real-time PCR.
Quantitative duplex real-time PCR (StepOnePlus; Applied Biosystems) was performed and multiplex PCR reactions were evaluated (TaqMan Gene Expression Master Mix, and TaqMan gene expression assays GAPDH [Hs99999905_m1] and MUC5AC [Hs00873651_mH]; Applied Biosystems) according to the manufacturer’s instructions. The following thermocycling conditions were used under standard mode as per manufacturers’ recommendations: 20 seconds at 95°C, followed by 40 cycles at 94°C for 1 second and 60°C for 20 seconds. Analysis and fold-change differences were determined by the comparative threshold cycle (CT) method, calculated from the ΔΔCT values using the formula 2−ΔΔCT. The MUC5AC transcript levels were normalized to a glyceraldehyde-3-phosphate dehydrogenase endogenous control. We used a human reference complementary DNA (Clontech Laboratories) as a calibrator to provide standardization across the PCR plates.25 The relative quantification was determined by using the standard curve method.
All data are presented as mean (SD). The prevalence of DED was calculated by sex, and the binomial distribution was used to estimate the corresponding 95% CIs. Depending on distribution of individual data, the significance of differences in mean MUC5AC concentration (nanograms per milligram) between groups was determined by the Steel test, the Kruskal-Wallis test, or the Wilcoxon rank sum test, and 95% CI was based on nonparametric Hodges-Lehmann estimates for the difference between 3 groups with a DED diagnosis, between 3 VDT working hours groups, or between symptomatic and asymptomatic groups. P values were determined using the Fisher exact test, an unpaired 2-tailed t test, the Steel test, the Kruskal-Wallis test, or the Wilcoxon rank sum test. All statistical analyses were performed using SAS, version 9.2 (SAS Institute Inc).
The participant characteristics and demographic data are reported in Table 1. The 96 individuals who volunteered for tear and impression cytology sampling included 60 men and 36 women aged 22 to 62 years. The mean age was 41.7 (9.8) years for all participants, with 44.3 (10.0) years for men and 37.4 (7.6) years for women. Participants between 30 and 49 years made up 61% of the total study population. The mean duration of VDT use was 8.2 (1.9) hours per day for all participants, with 7.7 (1.9) for men and 9.0 (1.6) for women.
The ocular surface findings and prevalence of DED are reported in Table 2 and Table 3, respectively. The mean Schirmer test I value was 19.4 (12.4) mm for all participants. Overall, 79% of the participants (n = 76) had Schirmer test I values greater than 5 mm. The mean TBUT was 3.8 (2.4) seconds for all participants. Participants with a TBUT greater than 5 seconds composed 18% (n = 17) of the study population. Most participants had fluorescein and lissamine green staining scores of less than 3 (92%  and 85% , respectively). The proportion of women (13.9%) with definite DED was 2 times greater than that of men (6.7%) and accounted for 9.4% of the study population. The proportion of men and women with probable DED was similar (53.3% and 63.9%, respectively) and accounted for 57.3% of the study population. There were no statistically significant differences in any of the test results (Schirmer test I, TBUT, and fluorescein and lissamine green staining scores) and DED prevalence between men and women.
The mean concentration of MUC5AC, normalized to the total amount of protein in tears, was 6.8 (7.5) ng/mg in 192 eyes. The mean MUC5AC concentration was 6.9 (7.8) ng/mg in men (n = 120) and 6.7 (7.3) ng/mg in women (n = 72). There was no significant difference in mean MUC5AC concentration between men and women (P = .83).
The mean tear MUC5AC concentrations and their relationship to DED diagnosis and VDT use are reported in Table 4. The mean MUC5AC concentration in the participants with definite DED (3.5 [3.2] ng/mg; 18 eyes) was lower than that in the participants without DED (8.2 [10.4] ng/mg; 64 eyes) (Steel test, P = .02), and the estimated difference was −2.17 (Hodges-Lehmann estimator, 95% CI, −4.67 to −0.30). The mean MUC5AC concentration in the group with long VDT working hours (>7 hours: 5.9 [6.1] ng/mg; 86 eyes) was lower than that in the group with short hours (<5 hours: 9.6 [12.3] ng/mg; 38 eyes) (Steel test, P = .049), and the estimated difference was −1.65 (95% CI, −3.12 to 0.00).
The relationship between MUC5AC concentration and ocular symptom frequency is reported in Table 5. According to our classification of DED symptom frequency, we found that the mean MUC5AC concentration was lower in the eye strain symptomatic group (5.1 [5.6] ng/mg; 86 eyes) than in the asymptomatic group (8.2 [9.4]; 106 eyes) (Wilcoxon rank sum test; P = .001), as well as lower in the excess tearing symptomatic group (3.4 [2.4]; 14 eyes) than in the asymptomatic group (7.1 [8.3]; 178 eyes) (Wilcoxon rank sum test; P = .04). The estimated differences in eye strain and excess tearing were −1.71 (95% CI, −2.86 to −0.63) and −1.58 (95% CI, −3.63 to −0.09), respectively. Regarding the dry sensation, the symptomatic group tended to have a lower MUC5AC concentration compared with the asymptomatic group (Wilcoxon rank sum test; P = .08). There were no significant MUC5AC concentration differences in relationship to the frequencies of other ocular symptoms between the symptomatic and asymptomatic groups.
The normalized level of MUC5AC messenger RNA (mRNA) expression in the conjunctiva of all participants with a sufficient amount for analysis was 1.00 (1.76) (187 eyes). There were no significant differences in conjunctival MUC5AC expression of participants with definite DED (1.03 [1.03]; 18 eyes) and probable DED (1.01 [2.00]; 110 eyes) compared with no DED (0.97 [1.47]; 59 eyes). In addition, the mean MUC5AC mRNA expression in the conjunctiva was similar to that in the eye strain symptomatic group (1.12 [2.31]; 86 eyes) as well as in the asymptomatic group (0.90 [1.11]; 101 eyes) of VDT users. There were no MUC5AC expression differences in the frequencies of other ocular symptoms between the symptomatic and asymptomatic groups.
Dry eye disease is an important public health problem that causes ocular discomfort, fatigue, and visual disturbances that may interfere with daily activities.26 On the basis of data from the largest epidemiologic studies concerning DED,27 including the Women’s Health Study,28 it has been estimated that approximately 3.23 million American women (7.8%) and 1.6 million men (4.7%) older than 50 years are affected by DED. Another questionnaire-based study in Japan8 showed that tens of millions of individuals have severe symptoms and probably a more episodic manifestation of the disease because of exposure to contributing factors such as VDTs. That questionnaire-based study reported that continuous exposure of office workers to VDTs for long periods of time is a risk factor for the development of DED; however, the investigators did not evaluate the relationship of subjective and objective tests of DED to changes in the molecular composition of the tear film in VDT users. In the present institutional study, we investigated the correlation between the duration of VDT work, the presence of symptoms and signs of DED, and the concentration of MUC5AC of office workers.
Mucins are heavily glycosylated glycoproteins, with carbohydrates composing 50% to 80% of their mass. A large gel-forming mucin, MUC5AC is secreted by goblet cells in the conjunctiva,11,29 moving throughout the tear film to flush out particulate matter through eyelid movement. The decrease in MUC5AC concentrations may impair the wettability of the ocular surface epithelium because of the hydrophilic nature of the secreted mucins, which results from the heavy glycosylation that helps to hold fluid on epithelial surfaces.13
In the present study, we report data indicating that a decrease in the MUC5AC concentration in tears may be one of the reasons why VDT users develop DED. First, we used the latest Japanese criteria for DED to categorize the participants and found a decrease in MUC5AC concentration in the tears of the definite DED group compared with the tears of the no DED group. A previous study14 found a decrease in the goblet cell–specific MUC5AC concentration in the tear fluid of patients with Sjögren syndrome who had symptoms of DED. In our study, participants in the definite DED group had severe symptoms of DED with clinical alterations of the ocular surface. Our study found a decreased MUC5AC concentration in the tear fluid of VDT users with definite DED who did not have severe hyperemia and epithelial disorders of the ocular surface. Second, we found a correlation between low MUC5AC concentration and prolonged VDT working hours. The use of VDTs is associated with a decreased frequency of blinking,30 excess evaporation of tear fluid,5 and hypofunction of the lacrimal glands,6 each of which can contribute to the pathogenesis of DED. Prolonged VDT use induces a decrease in tear secretion from the lacrimal glands6; however, it remains unknown whether the prolonged use induces a decrease in MUC5AC secretion from goblet cells in the human conjunctiva. Our data indicate that MUC5AC mRNA expression is not altered in VDT users. However, a previous study23 has shown that MUC5AC mRNA levels are significantly reduced in patients with moderate to severe DED. Because most of the definite DED in the present study was mild, we hypothesize that the difference in DED severity between the 2 studies accounts for the discrepancy in MUC5AC mRNA expression. Future research into the detailed mechanisms of MUC5AC regulation and secretion should help to elucidate the relationship between prolonged VDT use and decreased MUC5AC secretion.
Tear MUC5AC is primarily responsible for the rheologic properties of mucus secretions because of its hydrophilic character (which results from its heavy glycosylation) that helps hold fluid and lubricate the epithelial surface.31,32 We hypothesize that decreased tear MUC5AC concentrations in VDT users may easily be affected by several factors, such as decreased blinking and low humidity in computer environments. Interestingly, MUC5AC was lower in the eye strain symptomatic group (P < .001) than in the asymptomatic group despite our finding that prolonged daily VDT use was not always associated with eye strain (Supplement [eTable 3]).
The major limitation of the present study is the difference in sample numbers among the groups. For example, the definite DED group was smaller than the probable and no DED groups. In fact, the data were not normally distributed. We did not identify how much this imbalanced data size affected the final result. To compensate for the biased sample, a sample size larger than found in the present study is needed.
Consequently, in our study, the mean MUC5AC concentration in the definite DED group was lower than that in the no DED group, although the mean Schirmer test I value was similar between the groups. Therefore, our findings suggest that the MUC5AC concentration in tears may contribute to the stability of the human tear film33 and the low MUC5AC concentration in tears may increase stress and induce ocular discomfort.
Office workers with prolonged daily use of VDTs, as well as those with an increased frequency of eye strain, have significantly low MUC5AC concentrations in their tears. In addition, in the present study, MUC5AC concentration in the tears of patients with DED was lower than that in individuals without DED.
Submitted for Publication: May 21, 2013; final revision received February 10, 2014; accepted February 15, 2014.
Corresponding Author: Yuichi Uchino, MD, Department of Ophthalmology, School of Medicine, Keio University, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan (email@example.com).
Published Online: June 5, 2014. doi:10.1001/jamaophthalmol.2014.1008.
Author Contributions: Drs Y. Uchino and M. Uchino 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.
Study concept and design: Y. Uchino, M. Uchino, Yokoi, Dogru, Kinoshita, Tsubota.
Acquisition, analysis, or interpretation of data: Y. Uchino, M. Uchino, Yokoi, Dogru, Kawashima, Okada, Inaba, Tamaki, Komuro, Sonomura, Kato, Argüeso.
Drafting of the manuscript: Y. Uchino, M. Uchino, Yokoi, Dogru, Inaba, Tamaki.
Critical revision of the manuscript for important intellectual content: Y. Uchino, Dogru, Kawashima, Okada, Komuro, Sonomura, Kato, Argüeso, Kinoshita, Tsubota.
Statistical analysis: Y. Uchino.
Obtained funding: M. Uchino.
Administrative, technical, or material support: Y. Uchino, Yokoi, Okada.
Study supervision: Y. Uchino, M. Uchino, Dogru, Kawashima, Argüeso, Kinoshita, Tsubota.
Conflict of Interest Disclosures: Messrs Inaba and Tamaki are employees of Santen Pharmaceutical Co, Ltd. Drs Kinoshita and Tsubota are consultants for Santen Pharmaceutical Co, Ltd. No other disclosures were reported.
Funding/Support: This study was supported by Grant-in-Aid for Young Scientists (B) 2010, 2279192 from the Ministry of Health, Labour, and Welfare and the Ministry of Education, Science, Sports, and Culture. Provision of facilities, transport of equipment, data analysis, and data management were supported by Santen Pharmaceutical Co, Ltd.
Role of the Sponsor: The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.