Dynamic Change of Optical Quality in Patients With Dry Eye Disease

Abstract

Purpose.: To evaluate the dynamic changes of optical quality in patients with dry eye disease (DED) by using a double-pass system.

Methods.: Fifty-six patients with DED and 35 control subjects were included in this study. Each patient underwent an evaluation of the ocular surface, using the Ocular Surface Disease Index, tear film breakup time (TBUT), corneal and conjunctival staining (Oxford scale), and Schirmer I test. Using the tear film analysis program of the Optical Quality Analysis System, we recorded the objective scatter index (OSI) every 0.5 seconds within a 20-second period with the subjects asked to blink freely. Several parameters were established to evaluate the dynamic alterations of optical quality and the effects of blinks: OSI standard deviation (SD), ΔOSI, ΔOSI/time, blinking change (BC), blinking frequency (BF), and tolerant limitation (TL).

Results.: Patients with DED had significant alterations of optical quality compared to control subjects. The ΔOSI, ΔOSI/time, BC, and BF were significantly higher and TL was significantly lower in DED patients than controls. Furthermore, the ΔOSI and ΔOSI/time were also significantly higher in patients with severe DED than patients with mild disease. In univariate analysis, both TBUT and corneal staining were correlated with ΔOSI, ΔOSI/time, and BC. In multivariate analysis, ΔOSI/time was correlated to the corneal staining score.

Conclusions.: The double-pass system allowed the monitoring of the dynamic changes of optical quality in patients with DED. Optical quality should be evaluated in patients with DED because these alterations cannot be determined from the classical DED clinical tests.

Dry eye disease (DED) is a multifactorial disease of the tears and the ocular surface that results in symptoms of discomfort, visual disturbance, and tear film instability.^1,2 Although these symptoms are not often very severe, the visual disturbance and chronic ocular surface irritation observed in DED directly decrease quality of life and interfere with the ability to carry out daily functions such as driving, watching television, and using a computer.^3,4 The anterior corneal surface covered by the tear film is the most powerful refractive interface in the ocular system. Visual disturbance in DED may result from an abnormal or unstable precorneal tear film, which is supposed to form an intact coating on the corneal surface. The tear film consists of an aqueous gel with an outer lipid layer that retards the evaporation of tears between blinks.⁵ In dry eye patients, deficiency of tear secretion or alterations of tear film quality induce aberrations and scattering that directly interfere with the quality of the optical system. Tear film breakup contributes to the exposure of an irregular corneal epithelial surface and drastically increases intraocular scattering.⁶ Similarly, Tutt et al.⁷ have observed a combination of increased aberrations and scattering during the interblink period that degrades the quality of the image projected on the retina (retinal image quality, RIQ). To alleviate ocular symptoms and to obtain better image quality, a compensatory rise in the blinking rate appears in patients with dry eye to redistribute the tear film over the cornea more frequently. 

Tear film changes are dynamic, since the tear film is directly in contact with air and evaporates between blinks.⁸ Therefore, observing the real-time changes of the tear film might be a more direct and objective parameter in assessing dry eye. Several clinical studies have shown the variability of aberrations over time by using the Hartmann-Shack wavefront sensor.^9,10 Nevertheless, the wavefront sensor neglects the contribution of scattering in assessing RIQ and overestimates optical quality.^11,12 More recently, a novel method for evaluating RIQ has been developed, the double-pass (DP) system. This device uses the light projected into the ocular system through the pupil, recording RIQ after the light passes through the ocular media and is reflected by the retina. The DP image derives from the convolution of the point spread function from the first and second passes, and it is susceptible to both aberration and scattering.⁶ Previous studies have shown the usefulness of the DP system in monitoring optical quality in ocular diseases such as cataract and uveitis, and after LASIK.^13–19 Interestingly, the results of these studies suggest that the more distant the source of scattering from the retina, the more severe the alterations on the visual system.¹⁹ Since the tear film is the first ocular structure crossed by light, these results emphasize the considerable influence of tear film alterations on vision quality. Although the deterioration of RIQ due to tear film disruption in DED has been reported,⁶ there are no published data on the real dynamic changes of optical quality in patients with DED. The aim of the study was therefore to evaluate the dynamic changes of optical quality and blinking-related alterations induced by DED by using the real-time observation method with the DP system. 

Patients and Methods

Patients

A total of 56 patients with DED were recruited from the Cornea Unit of Beijing Tongren Hospital from May 2013 to July 2014. The diagnosis of DED was based on symptoms and the results of the following tests: the Ocular Surface Disease Index (OSDI), tear film breakup time (TBUT), corneal and conjunctival staining, and Schirmer I test.⁴ Tear film breakup time was measured by instilling fluorescein into the inferior palpebral conjunctiva and the test result was the average of three measurements. Then, corneal and conjunctival fluorescein staining were quantified by using the Oxford grading scheme. The Oxford score corresponded to the sum of the corneal and conjunctival (nasal and temporal) scores. The corneal staining score was limited to the evaluation of the corneal surface. The Schirmer I test was performed without anesthesia, and the patient's eyes were closed for 5 minutes before removing the test paper. All tests were performed by the same examiner (QL). 

Patients with DED were separated into two groups according to the severity-grading scheme of the 2007 International Dry Eye Workshop.^1,4 The severe dry eye group (28 patients) was defined as Schirmer I test ≤ 5 mm and TBUT ≤ 5 seconds. The mild dry eye group (28 patients) consisted of patients with Schirmer I test between 5 and 10 mm and/or TBUT between 5 and 10 seconds. Thirty-five age- and sex-matched control subjects were also recruited. All control subjects had no complaints of ocular surface symptoms and no abnormality on ocular surface tests. Exclusion criteria for all subjects were as follows: age below 18 years, history of ocular surgery or trauma, contact lenses wear within the past 3 months, best corrected visual acuity < 1.0, spherical error ≥ 6 diopters (D), cylindrical refractive error (>+0.5D or <−0.5D), and other ocular diseases or topical treatment that may influence quality of vision. 

The study protocol followed the tenets of the Declaration of Helsinki and was approved by the Institutional Review Board of Beijing Tongren Hospital. All subjects agreed with all the procedures and were informed in detail about the nature of this study. 

Double-Pass Instrument

Optical quality was evaluated by using the Optical Quality Analysis System (OQAS II; Visiometrics S.L., Tarrasa, Spain). The OQAS uses a laser diode with a wavelength of 780 nm instead of visible light as the source of light to avoid discomfort during the examination. Aberrations and intraocular scattering were evaluated by using the objective scatter index (OSI). It is defined as the ratio between the integrated light in the periphery ring and the central peak of the DP image and represents the impact on the DP image caused by aberration and scattering. Good repeatability of OSI has been reported by other investigators.¹² The “Tear Film Analysis” program included in the commercially available OQAS software was used to record dynamic changes of OSI values. This program consisted of a 20-second examination with OSI measurement every 0.5 second and it also provided a graph that showed the OSI change over time, with an average and standard deviation OSI value within the 20-second series (Fig. 1). 

Figure 1

 

Example of tear film analysis with OQAS. Objective scatter index values, blink marks, and their corresponding DP images during the analysis were recorded. The OQAS also provides a graph that shows the OSI change over a period of 20 seconds.

Double-Pass Parameter Acquisition

The optical quality changes were acquired before clinical ocular surface tests. The examination was performed in a room without illumination to obtain the largest pupil, with controlled temperature and humidity. The subject was asked to blink freely as usual during the double-pass analysis. All OQAS analyses were done by the same researcher who was masked to the results of clinical tests (C-HT). Several parameters were calculated from the OSI values acquired during the 20 seconds of dynamic tear film analysis. In this study, only the OSI values from the interblink period longer than 2 seconds were used for analysis (Fig. 2). The parameters were the following: OSI standard deviation (SD), ΔOSI, ΔOSI/time, blinking change (BC), blinking frequency (BF), and tolerant limitation (TL). Parameter definitions are shown in Table 1. 

Figure 2

 

Example of calculation for the OSI-related and blinking-related parameters. In this subject, SD, ΔOSI, and ΔOSI/time were calculated from three interblink periods and BC was calculated from three marked blinks. The values of the parameters described above were the following: SD = 0.200, ΔOSI = 0.510, ΔOSI/time = 0.130, BC = 0.450, BF = 6, and TL = 5.

Table 1

 

Definition of OSI-Related and Blinking-Related Parameters in Tear Film Analysis

 

Statistical Analysis

All values were expressed as mean ± SD. For each subject, both eyes were tested and one eye was selected randomly. A Kruskal-Wallis test was used to compare values between groups and the Spearman correlation test was used to evaluate the correlation between OQAS parameters and the results of dry eye tests. A multivariate linear regression analysis was used to investigate the relation between the correlated parameters. A P value < 0.05 was considered significant. All statistical analyses were done by using XLSTAT (version 2014.2.06; Addinsoft, Inc., Paris, France). 

Results

There was no statistical difference in age and sex between groups. For the dry eye clinical tests, both the severe and mild dry eye groups had lower TBUT and Schirmer test and higher OSDI, Oxford, and corneal staining scores than the control group. There were also statistical differences between the severe and mild dry eye groups for TBUT, Oxford score, corneal staining score, and Schirmer test. The clinical test results are provided in Table 2. 

Table 2

 

Demographic Information and Clinical Test Results

 

Compared with normal subjects, DED patients had significant changes in OSI-related parameters and blinking parameters. The ΔOSI, ΔOSI/time, BC, and BF were significantly higher in DED patients than controls. Conversely, the TL was significantly lower in DED patients than controls. Furthermore, the ΔOSI and ΔOSI/time were significantly higher in patients with severe DED than patients with mild disease. The analyses of OSI-related and blinking-related parameters are shown in Table 3 and Figures 3 and 4. 

Table 3

 

Analysis of OSI-Related and Blinking-Related Parameters

 

Figure 3

 

Results of OSI-related parameters: SD (A), ΔOSI (B), ΔOSI/time (C), and BC (D). Significant differences were observed between severe and mild DED and age-matched controls. *Analysis of variance, Kruskal-Wallis test, P < 0.05; **P < 0.01.

Figure 4

 

Results of blinking-related parameters: BF (A), TL (B). Significant differences were observed between severe and mild DED and age-matched controls. **Analysis of variance, Kruskal-Wallis test, P < 0.01.

In the DED groups, TBUT was correlated with Oxford score and corneal staining. Corneal staining was also correlated with sex, age, and Oxford score. There was no correlation between OSDI, TBUT, and Schirmer test. The results of clinical test correlations are shown in Table 4. 

Table 4

 

Results of Correlations Between Clinical Test Results in DED Patients (n = 56)

 

In univariate analysis, TBUT was correlated with ΔOSI, ΔOSI/time, and BC. Age and corneal staining were correlated with SD, ΔOSI, ΔOSI/time, and BC. The Oxford score was correlated with SD, BC, BF, and TL. Sex was correlated with ΔOSI/time. The results of clinical test correlations are shown in Table 5. Considering optical quality and blinking parameters, the ΔOSI was correlated with ΔOSI/time and BC. The ΔOSI/time was correlated with BC. Finally, the BF was correlated with TL. The results of OQAS parameter correlations are presented in Table 6. 

Table 5

 

Results of Correlations Between Clinical Test Results and OSI-/Blinking-Related Parameters in DED Patients (n = 56)

 

Table 6

 

Results of Correlations Between OSI-/Blinking-Related Parameters in DED Patients (n = 56)

 

In multivariate analysis, ΔOSI/time remained correlated with corneal staining (r = 0.318, P = 0.017). Age was correlated with SD (r = 0.407, P = 0.002), ΔOSI (r = 0.336, P = 0.011), and BC (r = 0.378, P = 0.004). The Oxford score remained correlated with BF (r = 0.437, P = 0.001) and TL (r = −0.281, P = 0.036). 

Discussion

Measured in a standard way, the visual acuity of patients with dry eye is often 20/20 but instability of the tear film induces aberrations^9,10 and scattering⁶ that directly decrease the quality of vision and consequently the quality of life of patients with DED. In clinical practice, the evaluation of dry eye severity is limited because there is a weak correlation between the different ocular surface tests and between these tests and the symptoms reported by patients.²⁰ The alterations of vision quality, not evaluated by standard clinical tests, could also explain the discrepancy between the symptoms and the results of clinical tests often observed in DED patients. Therefore, a more precise assessment of optical quality alterations and the related impact on quality of vision is important in DED patients. 

Previous studies have shown the ability of OQAS to evaluate the optical quality of eyes with DED or for assessing the lubricant effect of artificial eye drops.^6,13,21 However, all previous studies have evaluated optical quality changes by asking the subjects to keep their eyes open as much as possible, instead of observing the real optical quality changes when they blink as usual.^6,13,21 Interestingly, since the image used for the patient's fixation during OQAS examination is illuminated, our patients stated that they had similar feelings during the examination as when looking at a computer or cellular phone screen in their daily lives. In addition, to our knowledge, no study has yet assessed the usefulness of OSI-related parameters in evaluating optical quality in patients with different levels of dry eye severity. Our objective was therefore to develop several parameters related to the patient's real-life visual function, based on OSI. The SD represented the standard deviation of OSI values during interblink periods and might be a surrogate marker for vision fluctuations in dry eye patients. Similarly, the result of ΔOSI and ΔOSI/time showed that ocular scattering was increased and that this rise in OSI was faster in patients with dry eye than in normal subjects. Moreover, using these parameters, there were significant differences between mild and severe dry eye patients. We assume that these changes were due to tear film dynamic alterations, since opacities of the cornea, lens, or vitreous body do not vary in such a short period.²² 

Objective scatter index usually decreased after blinks, since the disrupted tear film is redistributed over the corneal surface. The BC represented the value of OSI changes between marked blinks. Although the BC was higher in patients with dry eye than in normal subjects, reflecting the poor quality of the tear film in these patients, it was noteworthy to observe that for some patients in the dry eye groups, some blinks were associated with an increase in OSI value (Fig. 3). A very abnormal tear film or secretions from the palpebral margin may lead to a poor blinking effect with increased vision quality alterations after the blink. This result might explain why some patients with DED have complaints of decreased vision after blinking. 

The OQAS also presented the ability to detect the blinking rate during 20 seconds of observation. Several studies have shown that the spontaneous blink rate (SBR) may be influenced by a reduction in optical quality, cognitive load, and ocular surface conditions or the environment.²³ Spontaneous blinking has been classified into two types: complete blinks and incomplete blinks.²⁴ The mechanical force produced by the contact between the eyelids and the ocular surface in complete blinks is thought to be the main driver for meibomian gland lipid spread into the tear film.²⁵ An incomplete blink may result in a prolonged exposure of the corneal surface and poor tear film quality, causing the poor blinking effect that was mentioned above. In the present study, OQAS recorded blinks when the upper eyelid blocked the incident beam, therefore including all complete blinks and most of the incomplete blinks. The BF reflecting the patient's ocular discomfort and irritation was thus correlated to Oxford score and corneal staining (Table 5). In addition, blinks divided the OSI analysis into several interblink periods. Since we observed that the frequency of SBR was directly influenced by ocular discomfort and subjective feelings of optical quality, we defined the longest interblink period as TL to evaluate how long the subjects usually stay without blinking in a normal condition of free blinking. This parameter also showed significant differences, with DED patients having shorter TL, in agreement with previous studies.²⁶ 

The parameters developed in the present study demonstrate a difference between normal subjects and dry eye patients and between different severity grades of DED. The SD, ΔOSI, ΔOSI/time, and BC were more related to optical quality, and BF and TL were associated with both optical quality and ocular discomfort or irritation. Consequently, in univariate analysis we found that the corneal staining score was correlated to optical quality parameters such as ΔOSI and ΔOSI/time, whereas the Oxford score (conjunctival and corneal scores) was correlated to BF and TL. Corneal epithelium alterations, by inducing ocular scattering, may affect vision quality more directly than conjunctival epithelium damage. Similarly, there were correlations between TBUT, corneal staining, and ΔOSI and ΔOSI/time. Considering the OQAS system, only the tear film located in front of the pupil area affects OSI-related parameters. Therefore, OSI parameters are more closely related to alterations of the tear film occurring in the cornea and particularly in the central cornea. Accordingly, in multivariate analysis, ΔOSI/time remained correlated to the corneal staining score. The OSDI reflects the subjective sensation, the Schirmer test is used for the quantification of tear secretion, and TBUT analyzes tear film breaks that can occur over the entire corneal surface. Conversely to OQAS parameters, none of these tests really evaluate the optical quality of tears in front of the visual axis. This may explain the absence of a direct correlation between OSI parameters and these clinical ocular surface tests, and thus emphasizes the role played by this test for specific optical quality evaluations in DED. 

Although the exploration with OQAS of the optical quality could be limited in elderly patients or patients with moderate or high astigmatism, this system with new OSI parameters could help monitor and record the dynamic change of the tear film and the visual quality alterations observed in dry eyes. Given that the calculation of these new parameters can be easily automated for clinical and research purposes, they might be useful new tools for the evaluation of optical quality in DED patients. 

Acknowledgments

The authors alone are responsible for the content and writing of the paper. 

Disclosure: C.-H. Tan, None; A. Labbé, None; Q. Liang, None; L. Qiao, None; C. Baudouin, None; X. Wan, None; N. Wang, None