Evaluation of the ocular surface in asymptomatic glaucoma patients on topical medications and following trabeculectomy - A cross-sectional study

Abstract

Purpose:

This present prospective, cross-sectional study aims to comprehensively evaluate the ocular surface in asymptomatic patients with diffuse blebs after trabeculectomy versus chronic anti-glaucoma medication use and compare it with the age-matched normal population.

Methods:

Objective clinical evaluation was done by tear film break-up time (TBUT) and Schirmer’s test (ST) in the three groups–trabeculectomy >6 months with a diffuse bleb (Wurzburg bleb classification score ≧10), chronic anti-glaucoma medication (AGM >6 months) group, and normal population. In all groups, tear film osmolarity was checked with the TearLab_® device (TearLab Corp., CA, USA), and subjective evaluation was performed by administering Ocular Surface Disease Index (OSDI) questionnaire. Patients already on chronic lubricants or any other drug for the treatment of dry eyes (viz. steroids, cyclosporin) or having symptoms suggestive of an abnormal ocular surface, who had undergone refractive or intraocular surgery, and contact lens users were excluded.

Results:

In total, 104 subjects/eyes were recruited over 6 weeks. Thirty-six eyes recruited in the trab group were compared with 33 eyes studied in the AGM group, and both these groups were compared to 35 normal eyes. When compared to normals, TBUT and ST were significantly lower (P = 0.003 and 0.014) and osmolarity and OSDI were statistically significantly higher (P = 0.007 and 0.003) in the AGM group, whereas only TBUT was statistically significantly different (P = 0.009) when the trab group was compared to normals. Also, when the trab group was compared to the AGM group, ST was found to be higher (P = 0.003) and osmolarity was lower (P = 0.034).

Conclusion:

To conclude, ocular surface is affected even in asymptomatic patients on AGM but near normalcy is possible following trabeculectomy when blebs are diffuse.

It is now well-established that chronic usage of topical anti-glaucoma medications (AGM) can cause changes in the ocular surface.[1,2] This has mostly been attributable to the preservatives in the medications, especially benzalkonium chloride (BAK). These preservatives prevent microbial infections by inhibiting microbes in the formulations; being cationic surfactants, they principally disrupt lipids membranes, including that of the tear film.[3] Disruption of tear film lipids leads to an unstable tear film causing evaporative dry eye. However, these topical AGMs are indispensable in the initial management of glaucoma for the reduction of intraocular pressure (IOP). To counter the effects of BAK, a limited number of medications with alternative preservatives or even preservative-free formulations are available. Nonetheless, surgical management may have to be deployed when there is intolerance to these AGM though surgery in glaucoma is usually undertaken when IOP is uncontrolled, or the disease is progressing. The most common glaucoma surgery performed is trabeculectomy (trab)-producing an extraocular drainage reservoir at the limbus (bleb). Over the last two decades, several other options have become available in the form of minimally invasive glaucoma surgery (MIGS),[4] these procedures and devices help to avoid a bleb formation, and are especially deployed in mild-to-moderate open-angle glaucoma. However, there is limited availability of MIGS in newly industrialized nations. Furthermore, the role of MIGS is constantly evolving but currently, such procedures are being performed predominantly with a view to the reduction in the burden of AGM. Worldwide, therefore, trabeculectomy remains the most performed glaucoma surgery. The presence of the bleb, especially when high, can impair corneal wetting and give rise to symptoms–the dysesthetic bleb. Neves et al.[5] reported that the height of the filtering bleb interferes with tear stability, and Lee et al.[6] demonstrated raised tear film osmolarity in post-trab patients with increased bleb height in approximately 40%. We, therefore, set out to comprehensively evaluate the ocular surface in asymptomatic patients who had undergone trabeculectomy (>6 months ago) and had diffuse low-lying blebs and compared them with patients who were on AGM >6 months and were also asymptomatic. These two groups were then compared with an age-matched normal group.

Methods

This was a prospective, single-center, cross-sectional study of consecutive patients presenting over 6 weeks. IRB approval was obtained from the Ethics Committee at the University and additionally, informed consent was obtained from all individuals recruited. Data handling was as per the tenets laid down by the Declaration of Helsinki. A detailed history was taken from all eligible patients initially, who then underwent a comprehensive ocular surface evaluation, and tests were administered in the same sequence in all the groups. A single observer administered all the tests to the patients to avoid inter-observer discrepancies. Three distinct groups were recruited.

Group 1–Trabeculectomy group–subjects in this group were recruited from a tertiary-level glaucoma practice who had undergone surgery in 2018 or earlier. At the time of recruitment, they were not on any AGM for >6 months. All those included had a diffuse low-lying bleb, as per inclusion criteria.

All trabeculectomy surgeries were performed by a single senior fellowship-trained senior glaucoma surgeon; all surgeries were performed in a repeatable fashion, described elsewhere in detail.[7] The technique as published is for combined glaucoma and cataract surgery through separate incisions; the procedure of trabeculectomy when performed in a separate site combined phaco-trab, is like that performed as stand-alone surgery. Usage of this technique afforded the creation of diffuse blebs. [Fig. 1-Left].

Figure 1.

Left: Diffuse bleb, scoring 11 points on the Wurzburg Bleb Classification System (see text) Right: Bleb height at the limbus (single arrowhead) is approximately one corneal thickness (double arrowhead)

Bleb morphology evaluation: All blebs were evaluated via the Wurzburg Bleb grading system,[8,9] by a senior glaucoma specialist. This was performed dynamically at the slit-lamp to check for the presence of microcysts and also for the determination of the bleb height vis-à-vis the corneal thickness. Only those blebs that scored above 10 (of a total of 12) and had a bleb height not greater than one corneal thickness[Fig. 1-Right.] were eligible for inclusion.

Group 2–Glaucoma patients on AGM >6 months.

Group 3–Healthy normals who presented to the eye clinic for refractive errors and ocular examination were within normal limits. Healthy volunteers were also included.

Exclusions: In all groups, subjects already on chronic lubricants or any other drug for the treatment of dry eyes (viz. steroids, cyclosporin) or having symptoms suggestive of the abnormal ocular surface, those who had undergone refractive surgery, and those who were contact lens users—were all excluded. Additionally, in the trab group, any bleb that did not have a diffuse morphology (i.e. were anterior, localized, and overhanging) and had a height > 1 corneal thickness, were excluded.

Subjective evaluation: This was performed by administering the Ocular Surface Disease Index^© (OSDI) questionnaire and the scores were compiled. The OSDI^© is assessed on a scale of 0 to 100, with higher scores representing greater disability.

Tear Film Osmolarity: This was obtained under standard conditions via the TearLab_® osmometer (CA, USA). The lab-on-a-chip system of the device consists of a disposable test card, which collects a very small tear sample with on-the-spot analysis and osmolarity values.

Objective evaluation: Clinical tests such as Schirmer’s test (ST) without anesthetic, Tear film break-up time (TBUT), and fluorescein stain pattern were administered by a single observer in all the groups.

ST was assessed after hooking the 5 mm folded end of an absorbent strip of paper over the margin of the lower lid in the lateral lower conjunctival sac, and measurement of the length of the wetting of the strip after 5 min (closed eyes) was recorded. The Schirmer tear test strip, which comprises absorbent strips of paper of size 35 mm × 5 mm was used.

Fluorescein strip (Fluorets, Chauvin Pharmaceuticals, France) was wetted with normal saline and was administered with topical anesthesia (Proparacaine HCl 0.5%, Sunways, IND) and TBUT was recorded as the time interval between the complete last blink and first appearance of discontinuity of tear film or “dry” spot on the cornea under cobalt blue filter of the slit lamp. It was recorded twice and the mean of the two values was taken.

Fluorescein stain (FS) was also administered to note the pattern of stain according to the National Eye Institute (NEI) fluorescein staining scale.[10] The cornea was divided into five zones and a standardized grading system of 0 to 3 was used for each of the five areas on each cornea. Grade 0 is specified when no staining is present, and the maximum score is 15.

Statistics: Descriptive statistics are presented as mean ± standard deviation (SD) along with 95% confidence intervals (CI). Univariate comparisons of continuous and quantitative inter-group variables were performed using the paired t-test analysis or the Wilcoxon signed-rank test and Chi-squared test for categorical variables. A P value of < 0.05 was considered statistically significant and all tests were two-tailed. All statistical analyses were performed using the IBM^® SPSS^® Statistics software v. 20 (Chicago, IL, USA).

Results

In total, 104 subjects/eyes were recruited over 6 weeks. Thirty-six eyes recruited in the trab group were compared with 33 eyes studied in the AGM group and both these groups were compared to 35 normal eyes. The age of patients in the trab group was (53.6 ± 17.7 years, 95% confidence interval [CI] [47.6, 59.6]), the AGM group was (59.1 ± 12.3 years, 95% CI [(54.7, 63.5]) and it was (53.7 ± 12.6 years, 95% CI [49.4, 58.0]) in the normal group; however, this was not statistically significant. Also, 65 males were recruited, and the male: female ratio did not differ between groups (P = 0.704).

In the AGM group, the mean number of medications being used was 2.3 ± 1.2, 95% CI (1.9, 2.7) for >6 months, with more than 65% being on two or more drugs. The AGM that were used included prostaglandin analogs (PGA), beta-blockers (BB), alpha adrenergics (AA), and carbonic anhydrase inhibitors (CAI) both dorzolamide and brinzolamide–either unfixed or in commercially available fixed drug combinations. The preservatives in these formulations were BAK, Purite, and SofZia. There was one eye that had two topical drugs–one AGM administered was “preservative-free” via the nouvelia bottle design–the other topical medication had BAK as a preservative. In the trab group, the mean duration from trab surgery was 21.4 ± 15.4 months, 95% CI (16.2, 26.6), which were on an average 3.2 ± 0.9, 95% CI [2.9, 3.5] number of AGM before trab surgery. At the time of recruitment, none of the eyes in the trab group were on AGM.

The severity of glaucoma was determined by visual fields criterion–mild-to-moderate mean deviation MD up to − 12.0 dB and severe was > −12.0 dB. In both the groups, n = 8 eyes were in the mild-to-moderate category—22.2% in the trab group and 24.2% in the AGM group.

Subjective evaluation OSDI score [Table 1]–Both groups of asymptomatic glaucoma patients scored higher compared with normals; however, the AGM group patients scored the highest (11.1 ± 11.7, 95% CI [6.9, 15.2])–this difference being significant (P = 0.003) when compared to normal. However, the total score in the trab group was not significant when compared to normal (P = 0.085) and also when compared to the AGM group (P = 0.854).

Table 1.

Ocular surface-related subjective and objective scores in the three groups-normal, anti-glaucoma medication (AGM) and trabeculectomy (Trab)

Parameter tested	Normal Group n=35	AGM group n=33	Trab group n=36	P Normal vs. AGM group	P Normal vs. Trab group	P Trab group vs. AGM
Ocular Surface Disease Index Score Mean±SD [95% CI]	4.7±10.1 [1.2, 8.2]	11.1±11.7 [6.9, 15.2]	7.5±12.3 [3.3, 11.7]	0.003	0.097	0.085
Osmolarity in mOsm/L Mean±SD [95% CI]	302.0±7.8 [299.3, 304.7]	309.0±12.9 [304.4, 313.6]	305.2±11.2 [301.4, 308.9]	0.007	0.401	0.034
Schirmer’s test in mm Mean±SD [95% CI]	18.1±7.7 [15.4, 20.7]	13.8±7.6 [11.1, 16.5]	20.3±9.1 [17.2, 23.4]	0.014	0.317	0.003
Tear film break-up time in seconds Mean±SD [95% CI]	6.4±2.3 [5.6, 7.2]	4.8±2.1 [4.0, 5.5]	5.0±2.2 [4.3, 5.7]	0.003	0.009	0.825
Grading of fluorescein stain Mean±SD [95% CI]	0.5±1.2 [0.1, 0.9]	2.8±2.4 [1.9, 3.6]	1.9±2.9 [0.9, 2.9]	<0.001	0.095	<0.001

Tear osmolarity [Table 1]–this parameter was significantly less (P = 0.034) when the trab group was compared to AGM, but not so when it was compared to normals (P = 0.401).

Objective Tests [Table 1].

ST without anesthesia–ST was significantly higher (P = 0.003) when trab group was compared to the AGM group but not so when compared to normal (P = 0.317).

TBUT–both glaucoma groups (AGM and Trab) had lower TBUT as compared with normals (P = 0.003 and 0.009, respectively) but did not differ between groups (P = 0.825)

Fluorescein stain pattern and grading, too, was higher in both the glaucoma groups, but significantly only when the AGM group was compared to normals (P < 0.001).

Discussion

Various authors have documented that the ocular surface can be affected by AGM and our study too has a similar outcome by virtue of all parameters (subjective and objective) being significantly different when the AGM group was compared to normals. However, notably, in our study subjects were asymptomatic and, to the best of our knowledge, changes in the ocular surface exclusively in such eyes have not been reported before. OSDI score as well as osmolarity and grading of fluorescein stain was higher in the AGM group and ST and TBUT were lower. However, when the trab group was compared with normals, OSDI, and osmolarity were higher but not significantly so. Even though surface wetting was slightly better in this group compared to normals (not significant), the reduction in TBUT was significant. This was, perhaps, testimony to the fact that the ocular surface is interrupted by the bleb; however, low and diffuse it may be, it can interfere with tear-film stability despite adequate production of tears. It is notable that Lee et al.[6] reported ST (without anesthesia) levels of less than 10 mm in their study comparing both types of glaucoma groups (AGM and Trab) with normals. In contrast, Neves et al.[5] reported levels of ST in the trab group and normals that reflect values we obtained; however, unlike their study, ST in the trab group (versus normals) did not reach significance in our study. The authors also reported significantly decreased TBUT and increased fluorescein stain, akin to our study. There is considerable variability in values obtained for tear production via ST by other authors too.[2,11,12] Yet, there appears to be a reluctance to discard this diagnostic parameter, perhaps because it is still one of the simplest, fastest, and least expensive diagnostic tests available. The greatest value of this test may be its ability to confirm which patient has severe dry eye when levels of 5 mm or lower are recorded.

With the creation of a bleb (and by default a “bump”) at the superior limbus, there is a risk of interference in the wetting of the surface as well as decreased TBUT, leading to dysesthetic symptoms. The greater the height of the bleb, the greater the hazard. Neves et al.[5] also reported that filtering blebs interfere with the stability of the ocular surface; their blebs were evaluated via the Indiana Bleb grading system, and 16 out of 40 blebs (40%) had greater height at the limbus, likely influencing their conclusions. Budenz et al.[13] have also conclusively proven that eyes with glaucoma filtering blebs experience more dysesthesia than eyes without filtering blebs. However, if a diffuse bleb can be created, with a height of no more than one corneal thickness (approximately 500 microns), then we have demonstrated that such eyes have an ocular surface, which almost resembles that of normals. We deliberately chose the Wurzburg bleb classification system (WBCS) for the evaluation of blebs in the inclusion criteria of the trab group, rather than alternate ones, as it lends an opportunity to recognize a functioning bleb without emphasizing the height of the bleb.[8] This latter feature is often, though sometimes erroneously, used as a surrogate feature in the determination of the functioning of a bleb. We believe, our reproducible technique of performing trabeculectomy,[7] aids the morphological evolution of such favorable blebs in the majority of cases. Also, with the technique, in approximately 80%, we are able to avoid the re-commencement of AGM well beyond 2 years.[7]

There are very few studies that have evaluated tear osmolarity in eyes with blebs as the TearLab_® osmometer is a relatively new technology for measuring the osmolarity of tears. It uses a micro-electrode to measure the number of charged particles in a tear sample without direct contact with the ocular surface, minimizing the chance of reflex tearing. Furthermore, it requires only a small tear sample (approximately 0.2 µL) and is able to provide results almost instantaneously. Evaluating the diagnostic ability of tear osmolarity requires specifying a value that discriminates a healthy eye from an eye with dry eye disease (DED). In a study, a tear osmolarity threshold of 308 mOsm/L diagnosed severe dry eye and normal patients 90.7% and 81.3% of the time, respectively; this value appeared to be the most sensitive for discriminating between normal eyes and those presenting with early stages of DED.[14] Currently, the 316 mOsm/L threshold is believed to better discriminate between mild and moderate/severe dry eye. A review by Potvin et al. appeared to support the use of tear osmolarity as a tool not only to diagnose and grade severity but also to track therapeutic response in DED.[15] In our study, we found that in eyes with chronic AGM, tear osmolarity was higher than the internationally accepted threshold of 308 mOsm/L and the levels in the trab group were closer to the normative group; this is in direct contrast to the report by Lee et al.[6] The presence of such differences in the two unrelated cohorts may exist due to differences in the use of anti-fibrotic agents, their dose, and duration during filtration surgery as well as the use of topical steroids in the postoperative period, morphology of the blebs at the time of recruitment, amongst other tangible and non-tangible factors. It is very evident from our data that tear osmolarity is significantly higher in AGM use and that the levels decrease significantly post-trabeculectomy.

Notably, all these ocular surface differences existed across groups, yet our patients were asymptomatic–an essential criterion for inclusion in the study. Surprisingly, even though all our subjects were asymptomatic, the OSDI_© scores were elevated across all groups, even in the normal group, but were significantly increased only in the AGM group when compared to normals. The OSDI_© is assessed on a scale of 0 to 100, with higher scores representing greater disability.[16] The index demonstrates sensitivity and specificity in distinguishing between normal subjects and patients with dry eyes. The index was the highest in the AGM group and intermediate in the trab group. It may be likely that the subjects were experiencing symptoms that they failed to attribute to either the AGM or the surgery. Furthermore, the influence of certain factors at the time of recruitment, principally environmental, cannot be excluded across all groups. Yet other factors that may be contributing could be anatomical in nature—position of the lids, inadequate closure during sleep, and floppy eyelids. Perhaps this may have influenced the increased subjective score in the normal group.

Our prospective observational study has several limitations. We concede that the surface changes may not be due to preservatives alone and that the active drug ingredient may also contribute, with loss of goblet cells in the conjunctiva seen in histopathology in various studies via impression cytology.[17] This was beyond the scope of our observational study. We also chose not to perform Rose Bengal staining (which highlights devitalized or dead cells, whereas fluorescein highlights epithelial cell loss) as it can cause severe irritation. Moreover, all subjects had self-reported as asymptomatic—it is possible that they may be experiencing symptoms but did not attribute it to the topical medications or the surgery they underwent. Furthermore, because we recruited only subjects with diffuse blebs (Wurzburg score >10) and a height not >1 corneal thickness, the results are not valid in those eyes with high and/or localized anterior blebs.

Conclusion

We conclude that the ocular surface is affected even in asymptomatic patients on AGM but near normalcy is possible following trabeculectomy when blebs are diffuse. Also, although glaucoma patients may be asymptomatic in terms of ocular surface disease, proactive assessment is recommended.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.