Dry eye parameters measured with an ocular surface analyzer in eyes after vitrectomy for vitreomacular interface disorders

Abstract

Purpose:

Dry eye disease (DED) might be caused by multiple ocular surgical interventions. The aim of the study was to estimate the extent of DED in patients undergoing core vitrectomy for vitreoretinal interface disorders.

Methods:

In this prospective observational study, we included patients with 12 months of follow-up after vitrectomy. The following data were collected as controls: age, sex, best-corrected visual acuity before and after surgery, and phakic status. In OSA (ocular surface analysis), the following parameters were evaluated: NIBUT (non-invasive tear break-up time), sltDear (thickness of the lipid layer), Meibomian gland (MGD) loss, and the height of tear meniscus. Shapiro–Wilk test, Wilcoxon rank-sum test, and Mann–Whitney U tests were used for statistical analysis.

Results:

We evaluated 48 eyes of 24 patients (10 men, 14 women; 64.63 ± 14.10 years) 1 year after vitrectomy. From the analyzed ocular surface parameters, NIBUT was significantly lower in operated versus non-operated eyes (P = 0.048). The higher the level of difference in MGD loss between both eyes, the higher the level of difference in NIBUT between both eyes (r_s = 0.47, P = 0.032).

Conclusion:

NIBUT levels were still decreased 12 months after vitrectomy. Patients with more pronounced MGD loss or decreased NIBUT levels in the fellow eye were more likely to experience such disorders. The tear meniscus height was lower in patients undergoing surgery for retinal detachment than in those with vitreoretinal disorders. This might allow the suggestion to include artificial tears in pre- and post-operative care in vitrectomized eyes.

Dry eye disease (DED) might be caused by multiple ocular surgical interventions, such as corneal refractive surgery, for example, laser-assisted in situ keratomileusis (LASIK) and keratoplasty but also cataract, glaucoma, and conjunctival surgery.[1] It was previously reported that DED after refractive surgery is either evaporative or because of obstructive Meibomian glands (MGDs).[2] There is also an additional neurotrophic component, which at least transiently compromises the function of the lacrimal functional unit.

In addition to slit-lamp ophthalmoscopy and vital dye staining, novel tools have recently been developed, which enable semi-automatic measurement of tear meniscus height, blinking characteristics, meibography, tear film interferometry,[3] or non-invasive tear break-up time (NIBUT).[4] OSA (ocular surface analysis) is a novel non-invasive, objective technique enabling the measurement of the following parameters: NIBUT, sltDear (thickness of the lipid layer), MGD loss, and height of the tear meniscus. This is the first study using OSA to measure the abovementioned DED parameters in patients 1 year after vitrectomy.

Methods

We prospectively evaluated 48 eyes of 24 patients (10 men, 14 women) 12 months after vitrectomy. The mean age of our patients was 64.63 ± 14.10. Twelve left eyes and 12 right eyes were operated on. The reasons for surgery were epiretinal membrane in seven cases, lamellar macular hole in two eyes, full-thickness macular hole in seven eyes, vitreomacular traction syndrome in two eyes, optic disc pit in one eye, vitreous hemorrhage in one eye, and retinal detachment in four eyes. At the end of the follow-up, six eyes were pseudophakic, and the remaining eyes were phakic.

This prospective analysis was approved by the local ethics committee and adhered to the tenets set forth in the Declaration of Helsinki. This cohort was composed of patients treated with vitrectomy with 12 months of follow-up. Both study and fellow eyes were evaluated. Fellow eyes were the control group. Inclusion criteria were as follows: unilateral vitrectomy in the anamnesis, with 12 months follow-up. The exclusion criteria were glaucoma, previous treatment with anti-vascular endothelial growth factor (anti-VEGF) injections, previous vitreoretinal surgery, uveitis, Sjogren disease and diabetes, and any ocular surgery prior to vitrectomy in either eye.

In all eyes, a complete ophthalmic examination and OSA with an ocular surface analyzer (SBM Sistemi Srl, Orbassano, Italy) were performed. None of the analyzed patients used artificial tears on a regular basis during the post-operative period. I.C.P. OSA enables to perform both quantitative and qualitative tests of the tear film, all of which are non-invasive. The patient is advised not to wear contact lenses on the day of the test and not to apply moisturizing eye drops shortly before the test. To make tests reliable, all of them are performed before pupil dilation.

The following data were collected: age, sex, best-corrected visual acuity before and after surgery, and phakic status. In OSA, the following parameters were evaluated: NIBUT, sltDear, MGD loss, and the height of the tear meniscus. All patients were evaluated 12 months after vitrectomy. NIBUT was measured without fluorescein dye after asking the patient to blink three consecutive times and then hold the eyes open. The normal NIBUT time was expected to be above 10 s. The measurement was repeated three times, and the mean value was registered [Fig. 1]. SltDear was estimated by observing the colors and interference pattern of the moving lipid tear film. Infrared meibography was performed after everting the lower eyelid. MGD loss was defined as the percentage of gland loss in relation to the total tarsal area of the eyelid. MGD loss was reported as moderate (25–50%) or severe (>50%) [Fig. 2]. The height of the tear meniscus was measured along the lower eyelid margin, immediately below the pupil. In the OSA diagnostic tool, a tear meniscus height above 0.2 mm is estimated as normal, 0.2–0.1 mm is a border outcome, and a result below 0.1 mm means significant tear deficiency [Fig. 3].

Figure 1.

NIBUT performed in vitrectomized eyes with OSA. (a) Correct outcome. (b) Incorrect outcome

Figure 2.

MGD loss presented by OSA in eyes 12 months after vitrectomy. (a) Photo of patients’ MGDs. (b) The loss area of MGDs is 27%

Figure 3.

Tear meniscus height measured with OSA in eyes 12 months after vitrectomy. (a) Border outcome. (b) Correct outcome

Statistical analysis was conducted in R software, version 3.5.1 (http://cran.r-project.org). Nominal variables are presented as n (%), and continuous variables are presented as the mean ± SD or median (Q1; Q3), depending on the data distribution. The normal distribution of data was checked with the Shapiro–Wilk test and based on skewness and kurtosis values. Comparisons of operated and non-operated eyes were conducted with the Wilcoxon rank-sum test or paired t-test. Analysis of independent groups was conducted with independent samples t-tests or Mann–Whitney U tests. Additionally, the mean (median) difference (MD) with a 95% confidence interval (CI) was calculated. For parameters that were significantly different between both eyes, Spearman’s correlation between differences and other parameters was calculated. All tests were two-tailed with a = 0.05.

Results

Twelve months after surgery, visual acuity was significantly lower in operated eyes (P < 0.001). The VA (visual acuity) level in operated eyes was 0.16 (0.09;0.30) versus 0.75 (0.50;0.98) for non-operated eyes, MD = -0.59; 95% CI [-0.75; -0.44], P < 0.001. From the analyzed ocular surface parameters, NIBUT was significantly lower in operated versus non-operated eyes (P = 0.048). The NIBUT level in operated eyes was 8.50 (7.55;9.28) versus 9.80 (8.23;10.25) for non-operated eyes, MD = -1.30; 95% CI [-2.00;-0.01], P = 0.048. For the remaining parameters, no significant differences were confirmed between operated and non-operated eyes [Table 1].

Table 1.

Analysis of operated vs. non-operated eye

Characteristic	Operated eye	Non-operated eye	MD (95% CI)	P
Final VA	0.16 (0.09;0.30)	0.75 (0.50;0.98)	-0.59 (-0.75;-0.44)	<0.0011
NIBUT	8.50 (7.55;9.28)	9.80 (8.23;10.25)	-1.30 (-2.00;-0.01)	0.0481
MGD loss (%)	10.54±8.02	7.96±6.58	2.58 (-1.02;6.18)	0.1512
Tear meniscus height	0.18 (0.17;0.27)	0.19 (0.15;0.22)	-0.01 (-0.06;0.01)	0.2181
sltDear	3.00 (2.00;3.00)	2.00 (2.00;3.00)	1.00 (-0.50;1.00)	0.3181

Data presented as mean±SD or median (Q1;Q3), depending on normality of data. MD – mean or median difference between eyes with 95% CI calculated as operated eye minus non-operated eye. Both eyes compared with Wilcoxon sum rank test¹ or paired t-test². sltDear expressed as numbers (D – 1, C – 2, B – 3, A – 4); N.I.B.U.T -non-invasive tear break-up time, VA, visual acuity; sltDear, thickness of the lipid layer

To identify parameters related to significant changes between both eyes, the correlations of NIBUT difference between both eyes with other parameters of both eyes were calculated. NIBUT difference was moderately negatively correlated with NIBUT level for the non-operated eye (r_s = -0.47, P = 0.026), meaning that the higher the level of NIBUT for the non-operated eye, the smaller the level of difference in NIBUT between the operated and non-operated eyes. Additionally, the NIBUT difference between eyes was moderately positively correlated with the difference in MGD loss between both eyes (r_s = 0.47, P = 0.032), meaning that the higher the level of difference in MGD loss between both eyes, the higher the level of difference in NIBUT between both eyes [Table 2].

Table 2.

Correlation between V-difference and NIBUT difference between both eyes with other parameters

Spearman’s correlation	VA difference		NIBUT difference
	r s	P	r s	P
Age, years	-0.05	0.827	0.37	0.086
VA of the operated eye	0.34	0.118	0.21	0.359
NIBUT of the operated eye	-0.17	0.449	0.35	0.110
MGD loss (%)	-0.13	0.555	0.29	0.187
Meniscus height of the operated eye	-0.05	0.796	-0.22	0.317
sltDear of the operated eye	0.01	0.978	-0.30	0.182
VA of the non-operated eye	-0.90	<0.001*	0.05	0.833
NIBUT of the non-operated eye	-0.13	0.555	-0.47	0.026*
MGD loss (%) of the non-operated eye	0.06	0.768	-0.08	0.714
Tear meniscus height of non-operated eye	-0.06	0.793	-0.22	0.319
sltDear of non-operated eye	0.11	0.609	-0.19	0.405
VA difference	-	-	-0.10	0.654
NIBUT difference	-0.10	0.654	-	-
MGD loss (%) difference	-0.30	0.155	0.46	0.032*
Tear meniscus height difference	-0.25	0.247	-0.06	0.790
sltDear difference	-0.24	0.266	-0.20	0.374

V difference, NIBUT difference and difference in other parameters calculated as operated eye minus non-operated eye. *P<0.05; NIBUT, non-invasive tear break up time; VA, visual acuity; sltDear, thickness of the lipid layer

There was no difference in any parameter between phakic and pseudophakic patients. NIBUT levels were significantly lower in the fellow eye in older patients (r_s = -0.52, P = 0.012); in vitrectomized eyes, this correlation was not significant (r_s = -0.37, P = 0.093). No statistical significance was found between MGD loss and the patient›s age or sex.

In addition, we analyzed particular parameters in regard to the type of vitreoretinal surgery. In fact, patients with retinal detachment had a significantly lower tear meniscus, which was 0.16 (0.12;0.18) versus 0.20 (0.14;0.53) in other patients, MD = 0.04; 95% CI [0.0001;0.15], P = 0.036. These values were insignificant for fellow eyes in those patients [Table 3].

Table 3.

Analysis of retinal detachment vs. non-retinal detachment patients

Characteristic	Retinal detachmenet patients (n=4)	Non-retinal detachment patients (n=18)	MD (95% CI)	P
Operated eye
Final VA	0.18 (0.01;0.30)	0.16 (0.00;1.00)	-0.02 (-0.14;0.24)	0.719
NIBUT	9.85 (8.30;13.30)	8.25 (4.60;11.60)	-1.60 (-4.70;0.40)	0.160
MGD loss (%)	8.50 (1.00;17.00)	7.50 (0.00;26.00)	-1.00 (-8.00;13.00)	0.609
Meniscus height of the operated eye	0.16 (0.12;0.18)	0.20 (0.14;0.53)	0.04 (0.0001;0.15)	0.036
sltDear	2,50 (2,00;4,00)	2,50 (2,00;4,00)	0.00 (-1.00;1.00)	0.706
Non-operated eye
Final VA	0.60 (0.30;1.00)	0.85 (0.20;4.00)	0.25 (-0.20;0.60)	0.366
NIBUT	10.80 (5.40;20.00)	9.40 (5.70;13.50)	-1.40 (-10.10;3.20)	0.418
MGD loss (%)	1.00 (0.00;9.00)	7.50 (0.00;25.00)	-2.50 (-0.01;14.00)	0.066
Meniscus height of the operated eye	0.15 (0.14;0.22)	0.20 (0.14;0.33)	0.05 (-0.01;0.08)	0.145
sltDear	3,00 (1,00;4,00)	2,00 (0,00;4,00)	-1.00 (-2.00;1.00)	0.492

Data presented as median (Q1;Q3). MD – median difference between groups with 95% CI calculated as non-OS minus OS. Both groups compared with Mann–Whitney U test. sltDear expressed as numbers (D – 1, C – 2, B – 3, A – 4); NIBUT, non-invasive tear break-up time; VA, visual acuity; sltDear, thickness of the lipid layer

Discussion

The current study presents the extent and type of DED occurring in patients after vitrectomy performed in non-diabetic patients. We observed a significantly lower NIBUT in operated versus non-operated eyes 12 months after surgery. Eyes with higher MGD loss were more likely to experience lower NIBUT levels. Eyes with lower NIBUT in the fellow eye usually experienced a more severe decrease in NIBUT in the vitrectomized eye. Patients after retinal detachment surgery had a lower tear meniscus than eyes operated on for vitreoretinal diseases. Tear Film and Ocular Surface Society Dry Eye WorkShop (TFOS DEWS) II describes DED as a multifactorial disease. The main etiopathogenetic factors are tear film instability, tear hyperosmolarity, inflammation, and neurosensory abnormalities.[3] As no gold standard diagnostic marker for DED has been established at this point, the diagnosis is based on anamnesis with the association of at least one marker of disrupted homeostasis of the ocular surface (corneal staining, tear film instability, or increased tear osmolarity).[5] Recently, a number of non-invasive tests have been developed to allow additional, automatic calculation of several parameters.[6] Among those, OSA enables the measurement of NIBUT, the thickness of the lipid layer, MGD loss, and height of the tear meniscus. In this study, we used this assessment tool to analyze the ocular surface in patients 1 year after vitrectomy.

Cataract surgery was reported to transiently (approximately 3 months) exacerbate the dry eye syndrome.[7,8] Most studies show a decrease in the tear break-up test (TBUT) and increased ocular surface staining in the first three post-operative months.[9] The pathophysiological mechanisms include exposure to topical anesthetics and light toxicity during surgery and additional nerve dissection.[10]

Only a few studies on the influence of vitrectomy are available, even though clinical practice experience shows that it is a quite frequent problem in such patients. During vitrectomy, the continuity of the conjunctiva is interrupted, which can, in turn, lead to DED. Heimann and co-workers presented that brachytherapy in retinal procedures resulted in 63% of patients experiencing dry eye symptoms.[11] Another study suggested that a 23 gauge vitrectomy might result in a lower rate of dry eye problems after 1 month after surgery. However, they performed a 360 peritomy during 20 G procedures, which is not usually done in such cases.[12] Thus, the aim of the current study is to estimate the extent of DED in patients undergoing core vitrectomy for vitreoretinal interface disorders 12 months after surgery.

It is well known that cataract or corneal surgery influences DED.[7,13] Even though clinical practice encounters a high frequency of DED in vitrectomized eyes, there is limited literature on that topic. Most studies report that up to 63% of patients experience dry eye symptoms and research is mainly focused on brachytherapy in retinal procedures.[11] Delayed wound healing is a widely known complication of diabetes. DED was previously reported in 50–60% of diabetic patients.[14,15] Corneal complications and delayed conjunctival wound healing were previously reported in diabetic vitrectomy.[16] However, even in this group, DED was not exhaustively evaluated.

It was previously discussed that trocars, as well as sutures, damage conjunctival and scleral structures. Heimann and co-workers noted an increase in conjunctival epithelial stratification, a decrease in PAS and MUC5AC-positive goblet cells, and distributional changes in MUC1, syndecan-1, and TN-C expression in conjunctival epithelium and stroma.[14] It was also suggested that during the first weeks after vitrectomy, the density of goblet cells was lowered.[17] Some authors suggest that there is no influence of small gauge vitrectomy on tear film stability.[18] Another study suggests that 3 months after vitrectomy, the tear break up was different between eyes with and without sutured sclerotomies. No differences were observed in the Schirmer test and corneal surface staining.[19]

In the current study, we confirm that even in the long term after vitrectomy, DED might be a significant problem for patients and might be more pronounced in older patients. Patients operated on for retinal detachment also had more severe DED symptoms than patients operated on for vitreoretinal interface diseases. This might be associated with more surgical maneuvers performed during retinal detachment surgery, for example, the possible use of ocular indentation, but also might be due to the length of surgery. Whether the patients were phakic or pseudophakic at the end of the follow-up did not further influence the ocular surface in vitrectomized eyes.

Future studies should focus on the evaluation of DED after vitrectomy at different time points.

Conclusion

In this study, we report that NIBUT levels are decreased in a long-term follow-up after vitrectomy. Patients with more pronounced MGD loss are more likely to experience such disorders. This might allow the suggestion to include artificial tears in post-operative care in vitrectomized eyes.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.