A New Noninvasive Tear Stability Analysis System for the Assessment of Dry Eyes

Abstract

purpose. This was a prospective case–control study conducted to evaluate the effectiveness of the Tear Stability Analysis System (TSAS) for the assessment of patients with dry eye.

methods. The TSAS can take 10 consecutive corneal topograms at one per second for 10 seconds. Examinations using the TSAS were conducted in 26 eyes of 26 healthy control subjects and in 27 eyes of 27 patients with dry eye. Examinations were also conducted in 14 eyes of 14 patients before and after the insertion of punctum plugs. Surface regularity and asymmetry indices (SRI, SAI), as well as new tear stability regularity and asymmetry indices (TSRI, TSAI), derived from SRI and SAI, were analyzed.

results. The mean SRI and SAI in dry eyes were significantly greater than in control eyes (P < 0.05). The time-wise change of SRI and SAI was significantly different between dry eyes and control eyes (P < 0.05). TSRI and TSAI in dry eyes were also significantly greater than in control eyes. Punctum plug insertion was associated with a significant decrease in SRI and SAI (P < 0.05).

conclusions. TSAS was effective in objectively assessing the tear stability in patients with dry eye. This system may be useful in noninvasive diagnosis of dry eye and evaluation of treatment effects.

Astable and continuous tear film is essential for achieving an optically smooth surface. Prolonged gaze without blinking results in tear film instability, which has been reported to be associated with decreased functional visual acuity and worsening of the surface regularity index (SRI) in conventional corneal topography in patients with dry eye. ¹  

The stability of tear film depends on many factors, such as intact blink reflex mechanisms, presence of healthy lacrimal glandular tissue, and structurally intact tear film. All three layers of the tear film—namely, the mucin, aqueous, and lipid layers—significantly contribute to tear stability. Qualitative and quantitative disorders of any of these layers can affect the tear film stability immensely. 

Tear film stability can be analyzed by invasive techniques that use fluorescein, such as tear film break-up time (BUT) and tear clearance tests, ² as well as noninvasive methods including tear film lipid layer interferometry, ^{3 4} tear evaporation test, and BUT assessment by using a grid xeroscope. ⁵  

BUT analysis with fluorescein dye is the most commonly used test of tear stability. Although it is easy to perform, variations in the concentration and pH of the fluorescein solution, the volume of the instilled drops, the presence of preservatives and invasiveness of the procedure are the main sources of error in this method. Research into noninvasive methods resulted in the development of techniques for the assessment of precorneal tear film BUT without the use of fluorescein (noninvasive BUT). Noninvasive BUT using instruments such as the grid xeroscope or tearscope (Keeler, Windsor, UK) allowed evaluation of the tear film by eliminating physical disturbance of the film from the instillation of fluorescein, along with the possibility of reflex tearing. Noninvasive BUT has been reported to be valuable for the diagnosis of tear film mucous layer deficiencies. However, noninvasive BUT did not find widespread acceptance in clinical practice due to problems in quantification of tear film stability. 

Corneal topographical examinations have also been shown to help in the assessment of corneal surface irregularities in aqueous tear deficiency states, by using indices such as SRI and SAI. ^{6 7} These indices represent local variations in corneal contour providing invaluable information about the relation of the corneal and tear film status with progression of corneal disease and irregular astigmatism. However, conventional corneal topography examination methods can provide corneal surface data at only one time point. We developed a software program and called it the Tear Stability Analysis System (TSAS) for the TMS-2N corneal topography instrument (Tomey Technology, Nagoya, Japan), which can take 10 consecutive corneal topograms, one per second for 10 seconds. TSAS can detect subtle time-wise changes in the tear film deriving data from the distortion of the mire rings. Goto et al. ⁸ recently reported that TSAS could detect subtle tear film instability in eyes with normal BUT, by using fluorescein dye. 

In this study, we performed TSAS measurements in patients with dry eye, to evaluate the effectiveness of this new system in the diagnosis of tear film instability, and compared the results with those in normal control subjects. We also assessed the changes in tear stability in subjects with dry eye who were treated with punctum plugs, to investigate the applicability of TSAS in the evaluation of dry eye management with punctal occlusion. 

Methods

Monocular TSAS examinations were performed on 17 right eyes of 17 patients with non-Sjögren syndrome (non-SS; 5 men and 12 women; average age, 57.8 ± 8.4 years) and 10 eyes of 10 patients with Sjögren’s syndrome (SS; 3 men and 7 women; average age, 62.1 ± 7.2 years). Twenty-six right eyes of 26 healthy control subjects also underwent the same examinations (8 men and 18 women; average age, 56.4 ± 9.4 years). There were no statistically significant age and sex differences between the control subjects and patients in this study. The research adhered to the tenets of the Declaration of Helsinki, and informed consent was obtained from all the subjects after explanation of the nature and possible consequences of the study. 

Tear Function Diagnostic Criteria and Ocular Surface Evaluations

The diagnosis of dry eye was based on the diagnostic criteria of the Dry Eye Research Group in Japan. ^{9 10} In brief, patients with dry eye-related symptoms, positive staining with fluorescein or rose bengal, and Schirmer 1 test results of less than 5 mm or tear BUT of less than 5 seconds were diagnosed to have dry eye. The ocular surface was examined by the double vital staining method. ¹¹ Two microliters of preservative-free combination of 1% rose bengal and 1% fluorescein dye was instilled in the conjunctival sac. Rose bengal staining of the ocular surface was scored according to the criteria proposed by van Bijsterveld ¹² and fluorescein staining was scored according to the protocol described by Shimmura et al. ¹³ A fluorescein staining score above 1 point and a rose bengal staining score of more than 3 points was considered abnormal. BUT was measured three times, and the mean value was calculated. ¹¹ Dry eye cases were categorized as non-SS and SS on the basis of the criteria proposed by Fox et al. ¹⁴ We recruited healthy control subjects with normal tear function and no vital staining of the ocular surface. The patients and control subjects did not have any history of ocular surgery, including punctal occlusion, ocular or systemic disease, or a history of drug or contact lens use that would alter the ocular surface. TSAS measurements were conducted before the tear function and vital staining examinations to minimize their effects on the TSAS testings. According to the study protocol, tear film BUT analysis was performed afterward, followed by fluorescein and rose bengal vital staining of the ocular surface. Schirmer I test was performed finally 1 hour after the initial TSAS testing to minimize the effect of the application of local anesthetic. 

TSAS Measurements

We developed a new software called TSAS in collaboration with Department of Ophthalmology, Ehime University and Tomey Technology (Nagoya, Japan). TSAS was installed in the Topographic Modeling System (TMS-2N; Computed Anatomy, Tomey Technology). In TSAS testings, 30 mire rings are projected on the corneal surface. The projected images are captured by videokeratoscopy. The ring data are decomposed into 256 individual point data and analyzed by computer. Briefly, the center of the innermost mire is established as a reference point during testing. Having established a central reference point, coordinates are given to each data point where a semimeridian intersects a mire. The 256 equally spaced semimeridians theoretically provide approximately 6,000 to 11,000 data points. The actual number of data points available for analysis maybe reduced by mire distortion induced by shadows, position of the eyelids, corneal surface, and tear film irregularities. The coordinates locating the data points are converted to polar coordinates on the keratoscopy mires to facilitate corneal topographical reconstruction. A reconstruction algorithm is then applied to the location of each point on the two dimensional reflection. The algorithms reflect the shape of the cornea and are presented as statistical indices in topographic displays. Statistical indices are numbers that summarize a particular feature of the cornea. Among them, the surface regularity index (SRI) is the measure of the local regularity of the corneal surface within the central 4.5-mm diameter. ⁷ Within this area, the power of each point is compared with that of the points immediately surrounding it. This index has been reported to correlate well with visual function. The construction of the SRI algorithm is shown in 1 . In contrast, the surface asymmetry index (SAI) is the measure of the difference in corneal powers between points on the same ring 180° apart. ¹⁵ The power distribution across a normal corneal surface is fairly symmetric, and this index has been reported to increase in corneal disorders indicating progression of the disease state. 2 shows construction of the SAI algorithm in detail. During TSAS measurements, corneal topograms were taken every second for 10 seconds after opening the eyes. To be able to perform the TSAS measurement without blinking, we applied 30 μL of 0.4% oxybuprocaine to the eyes to anesthetize the ocular surface. After 5 minutes, patients and normal control subjects were asked to stop blinking and keep their eyes open while their tear stability was being measured. The same experienced examiner performed all TSAS measurements obtaining well-focused and properly aligned images of the eyes. Eleven consecutive images were displayed on one printout. Mean surface regularity (SRI) and surface asymmetry index (SAI) information was provided in the upper left and right areas of each image. The printouts also provided information on the time-wise transition of SRI and SAI. In this study, we defined two new tear stability indices, which were calculated as follows: (1) TSRI = maximum SRI − minimum SRI within 10 seconds; (2) TSAI = maximum SAI − minimum SAI within 10 seconds. 

Kinetic measurements of ocular surface regularity after 10 seconds of sustained eye opening were performed in patients with dry eye and in normal control subjects. SRI, SAI, TSRI, and TSAI were compared between these groups. Likewise, the same comparisons were performed between the SS and non-SS groups, including comparisons before and after plug insertion. 

Punctum Plug Procedure

Punctum plugs were inserted in 14 eyes of 14 patients with dry eye. Two types of punctum plugs (FCI plug; FCI Ophthalmics, Issy-Les-Moulineaux, France, and Eagleplug; Eagle Vision, Memphis, TN) were used in this study. Plugs were inserted in both superior and inferior lacrimal puncta. Tear function and ocular surface as well as TSAS measurements were performed before and 2 weeks after plug insertion. 

Statistical Analyses

A two-way repeated-measures ANOVA was performed for the comparison of tear film and corneal surface regularity, as well as asymmetry index differences between the dry eye and control groups. The same test was instituted to test the index differences before and after the punctum plug treatment. The two- way repeated- measures ANOVA was used to assess the time-wise differences of surface irregularity and asymmetry indices with TSAS measurements. Student’s t-test was used for the comparison of TSRI and TSAI. A χ² test was applied to assess the age and sex differences between the control subjects and patients. P < 0.05 was considered statistically significant. A computer was used for statistical analysis (StatView software Windows 98/2000; SAS Institute Inc., Cary, NC). 

Results

Tear Function and Tear Stability Assessment of Patients with Dry Eye and Normal Subjects

The tear functions and vital staining scores of the subjects with dry eye and the control subjects are shown in 1 . TSAS findings in a representative control subject with no considerable changes in the regularity and asymmetry indices are shown in 3 . 4 demonstrates the gradual increase of the same indices in a patient with SS who had severe dry eye. TSAS examination showed that mean SRI and SAI from 0 to 10 seconds were significantly higher in the patients with dry eye than in the normal subjects (P < 0.05; 5 6 ). The time-wise variations of SRI and SAI were also statistically significant within each group (P < 0.05). The TSRI and TSAI in the patients with dry eye were also significantly higher than in the normal control subjects (2 ; P < 0.05). 

Comparison of Tear Functions and Tear Stability between the SS and Non-SS Groups

The tear functions and vital staining scores of the SS and non-SS dry eye groups are shown in 1 . TSAS examination showed that mean SRI and SAI from 0 to 10 seconds were significantly higher in patients with SS than in patients with non-SS dry eye (P < 0.05; 7 8 ). The time-wise variation of the SRI from 0 to 6 seconds was observed to be significantly different between the SS and non-SS groups (7 ; P < 0.05). Likewise, the time-wise variation of the SAI from 0 to 10 seconds were observed to be significantly different between the SS and non-SS groups (8 ; P < 0.05). 

Comparison of Tear Function and Tear Stability before and after Punctum Plug Insertion

We observed no complications related to the punctum plug insertion. The tear functions and vital staining scores of the patients with dry eye before and after plug insertion are shown in 1 . After punctum plug insertion, mean SRI and SAI from 0 to 10 seconds showed a significant improvement compared with the pretreatment indices (P < 0.05; 9 10 ). The time-wise variations of SRI and SAI were not significantly different (P > 0.05). 

Discussion

The efficacy of the TSAS, using videokeratography as a noninvasive and an objective method of tear stability assessment, has already been established. ⁸ The only available study on tear film stability employing this new device analyzed tear film stability using parameters such as BUT and break-up area. ⁸ In this study, we used the TSAS to measure the kinetic tear stability changes in patients with dry eye and normal subjects by using SRI, SAI, and two new indices, TSAI and TSRI; analyzed the tear stability index differences between patients with SS and those with non-SS dry eye; and looked into the changes in stability indices with punctum plug treatment in patients with dry eye. We have reported that functional visual acuity in patients with dry eye significantly decrease with concomitant increase in SRI. ¹ Decreased functional visual acuity in that study suggests impaired visual function on prolonged gaze in subjects with dry eye. 

The new TSAS allowed evaluation of the kinetic tear stability and corneal topographical changes during 10 seconds of blink-free gaze. TSAS assesses the changes of topographical mires reflected from the tear film surface. Although it is still not clear whether the tear stability changes measured by TSAS are due to complete full-thickness tear film disruption or discontinuity of the tear lipid layer, further studies in which TSAS and lipid layer interferometry are performed simultaneously on the same subset of patients with dry eye and control subjects will clarify these issues. TSAS examinations showed that the kinetic tear stability as assessed by SRI and SAI was significantly worse in patients with dry eye. Likewise, TSRI and TSAI fared significantly worse in the subjects with dry eye. Our previous attempts to performed TSAS measurements without topical anesthesia resulted in reflex tearing in some patients with dry eye who could not tolerate long blink-free periods. Thus, we used topical anesthesia in this study. It has been confirmed that a single application of oxybuprocaine does not affect the visual acuity or tear film stability itself in patients with dry eye or normal control subjects. ^{16 17 18}  

We were surprised to find that the kinetic tear stability patterns in patients with SS or non-SS dry eye were different from each other. The SRI and SAI indices were significantly lower initially in the non-SS subjects, in whom they increased within the first few seconds and displayed values closer to those in the patients with SS afterward. We believe that comparatively poorer tearing and a higher degree of corneal epithelial damage were responsible for the higher initial and consistently higher SRI and SAI scores in the patients with SS. Patients with non-SS dry eye had relatively better tearing and lesser vital staining scores with better kinetic stability, as evidenced by lower initial SRI and SAI scores, which worsened gradually after prolonged gaze without blinking. Although we did not quantify them in this study, we attributed the worsening of the surface regularity and asymmetry indices to the changes in tear evaporation. We observed that the magnitude of kinetic tear stability changes were much less in patients with SS compared with patients with non-SS dry eye. Thus, the new TSRI and TSAI were thought to be effective in the detection of the tear stability changes in patients with mild dry eye with low rose bengal and fluorescein staining scores. TSAS examination was also useful for the evaluation of dry eye management with punctum plugs. We expected a worsening of the pretreatment SRI and SAI in the patients with dry eye selected for punctal occlusion. However, the time-wise variation of the indices before occlusion did not show a marked change, probably because patients who underwent punctum plug treatment had considerably higher ocular surface vital staining scores than the overall group with dry eye in this study and because the SRI and SAI did not get worse during the testing. The SRI and SAI improved significantly with punctal occlusion, which suggests attainment of tear stability with treatment. 

In conclusion, TSAS helped in the assessment of tear stability in patients with dry eye and seemed to be promising in differentiating types of dry eye by quantitatively evaluating the dynamic changes of tear stability over 10 seconds. TSAS was also useful in evaluating the effect of punctal occlusion therapy. We believe that studies with larger number of subjects that look into the correlation of tear clearance rate with the tear stability indices would be very interesting. Measurements of tear stability within shorter intervals will increase the sensitivity of noninvasive tear film stability evaluations by TSAS in the near future. 

 Presented in part at the annual meeting of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, May 2003, and at the 27th Japan Cornea Congress, Karuizawa, Japan, February 17–19, 2003.

 Supported by grants from the Japan Ministry of Education and Science, Tokyo, and the Hightech Research Center, Tokyo Dental College, Chiba, Japan.

 Submitted for publication July 10, 2003; revised October 8 and November 19, 2003; accepted November 21, 2003.

 Disclosure: T. Kojima, None; R. Ishida, None; M. Dogru, None; E. Goto, None; Y. Takano, None; Y. Matsumoto, None; M. Kaido, None; Y. Ohashi, None; K. Tsubota, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Figure 1.

 

The definition of SRI. A, B, scaling constants; P_i,j, the matrix of corneal power; i, the hemimeridian index; j, the mire ring number; N, total number of powers within the ±1.0-D window.

Figure 2.

 

The definition of SAI. Symbols are as in 1 , with the addition of F_ci, which is a scaling constant corresponding to the mire ring.

Table 1.

 

 

Tear Function and Vital Staining Scores

Table 1.

 

Tear Function and Vital Staining Scores

	Schirmer (mm)	BUT (sec)	FL	RB
Dry eye and control
Dry eye (n = 27)	7.5 ± 4.6	2.6 ± 1.2	3.8 ± 2.7	4.6 ± 2.9
Control (n = 26)	14.3 ± 3.8	12.8 ± 3.4	0.0 ± 0.0	0.0 ± 0.0
Non-SS and SS
Non-SS (n = 17)	10.5 ± 5.9	2.5 ± 1.3	2.9 ± 2.3	3.7 ± 2.6
SS (n = 10)	4.8 ± 3.1	1.7 ± 0.8	5.0 ± 2.1	5.8 ± 3.1
Before and after punctum plug
Before plug (n = 14)	5.7 ± 2.7	2.3 ± 1.2	4.4 ± 2.7	5.3 ± 2.8
After plug (n = 14)	8.3 ± 3.3	3.6 ± 1.5	1.8 ± 2.3	2.6 ± 1.2

 Data provided as (mean ± SD).

 FL, score by fluorescein dye; RB, score by rose bengal dye; non-SS, non-Sjögren syndrome; SS, Sjögren syndrome; Before plug, before punctal plug insertion; After plug, after punctal plug insertion.

Figure 3.

 

Representative TSAS map from a normal subject, a 38-year-old man who had a Schirmer test result of 10 mm. BUT was 12 seconds, and there was no vital staining score. Note the stability of the map as well as the SRI and SAI. Inset: SRI (blue) and SAI (red) transition with the eye open.

Figure 4.

 

A representative TSAS map of a patient with dry eye, a 57-year-old man who had a Schirmer test result of 2 mm and BUT of 3 seconds. Rose bengal and fluorescein scores were both 6. Note the dramatic change in the TSAS pattern with time. Inset: the SRI (blue) and SAI (red) increased with time with the eye open.

Figure 5.

 

Comparison of SRI between patients with dry eye and control subjects. Error bars: standard deviation. The mean SRI from 0 to 10 seconds was significantly higher in the patients with dry eye than in the normal subjects (P < 0.05). The time-wise variations of SRI were also statistically significant within each group (P < 0.05).

Figure 6.

 

Comparison of SAI between patients with dry eye and control subjects. Error bars: standard deviation. The mean SAI from 0 to 10 seconds was significantly higher in the patients with dry eye than in the normal subjects (P < 0.05). The time-wise variations of SAI were also statistically significant within each group (P < 0.05).

Table 2.

 

 

TSI of Patients with Dry Eye and Normal Control Subjects

Table 2.

 

TSI of Patients with Dry Eye and Normal Control Subjects

Indices	TSRI	TSAI
Dry eye (n = 27)	1.3 ± 0.4*	2.1 ± 1.3*
Normal control (n = 26)	0.72 ± 0.3	1.1 ± 0.9

 Data provided as (mean ± SD). TSRI, Tear Stability Regularity Index; TSAI, Tear Stability Asymmetry Index.

* P < 0.05.

Figure 7.

 

Comparison of SRI between the SS and non-SS groups. Error bars: standard deviation. The mean SRI from 0 to 10 seconds was significantly higher in the SS than in the non-SS group (P < 0.05). The time-wise variation of the SRI from 0 to 6 seconds was significantly different between the SS and non-SS groups as shown (P < 0.05).

Figure 8.

 

Comparison of SAI between SS and non-SS groups. Error bars: standard deviation. The mean SAI from 0 to 10 seconds was significantly higher in the SS than in the non-SS group (P < 0.05). The time-wise variation of the SAI from 0 to 10 seconds was significantly different between the SS and non-SS groups (P < 0.05).

Figure 9.

 

Comparison of SRI before and after punctum plug insertion. Error bars: standard deviation. The mean SRI from 0 to 10 seconds showed a significant improvement compared with the pretreatment indices (P < 0.05). The time-wise variations of SRI were not significantly different (P > 0.05).

Figure 10.

 

Comparison of SAI before and after punctum plug insertion. Error bars: standard deviation. Mean SAI from 0 to 10 seconds showed a significant improvement compared with the pretreatment indices (P < 0.05). The time-wise variations of SAI were not significantly different (P > 0.05).