Morphological and Functional Assessment of Meibomian Glands in Keratoconus

Purpose: To evaluate the morphology and function of the meibomian glands (MG) in keratoconus patients. Methods: One hundred eyes of 100 keratoconus patients and 100 eyes of 100 age-matched healthy subjects were included into this study. Ocular Surface Disease Index (OSDI) scores, non-invasive break up time (NIBUT), meibography ndings, uorescein staining of the ocular surface, tear lm break-up time (TBUT), and Schirmer I test were recorded in all patient eyes and were compared with controls. Results: The mean TBUT and NIBUT were signicantly lower, corneal staining and OSDI scores were statistically higher in the keratoconus group (p<0.05). The mean meiboscore, partial gland, gland dropout and gland thickening scores of the upper and lower eyelids were signicantly higher in keratoconus patients compared to controls (p<0.05). The NIBUT measurements signicantly correlated with MG loss in both upper and lower eyelids (p<0.05). The severity of keratoconus seemed to correlate with meiboscore, partial gland, gland thickening scores in both upper and lower eyelids. Conclusion: Our data suggest that corneal ectasia in keratoconus is associated with alterations in ocular surface, tear lm function and MG morphology. Early screening and treatment of MG dysfunction may improve ocular surface quality and allow better disease management in keratoconus patients.


Introduction
Keratoconus is a progressive, bilateral, ectatic disease of the central and paracentral cornea characterized by stromal thinning, irregular corneal astigmatism, and myopia. The etiology of the disease remains unknown, however, possible linkage with systemic diseases, a combination of genetic predisposition and environmental factors such as excessive eye rubbing and use of contact lenses, have been proposed as possible mechanisms.
Although keratoconus has long been described as a non-in ammatory disorder, many studies have pointed to the possibility of underlying in ammatory mechanism which may link keratoconus to tear lm and ocular surface dysfunction [1,2]. Tissue degradation in keratoconus, involves the expression of in ammatory mediators, such as proin ammatory cytokines, cell adhesion molecules, and matrix metalloproteinases (MMP). Overexpression of interleukin (IL)-6 and tumor necrosis factor-alpha (TNF-α), MMP-9 was found in the tear lm of keratoconic eyes when compared to normal [3]. Eye rubbing, which is very common in keratoconus patients also results in an additional increase in levels of MMP-13, IL-6, and TNF-α levels in the tear lm [4].
Recently, there has been an increased interest in ocular surface and tear lm alterations of keratoconus.
Ocular surface damage with corneal and conjunctival staining, tear lm instability, and squamous metaplasia with goblet cell loss and decreased mucin levels have been demonstrated in keratoconic eyes in few studies [5][6][7]. Fukuchi et al demonstrated that the levels of lysosomal hydrolases are elevated in both conjunctival and corneal epithelium of patients with keratoconus, suggesting that the whole ocular surface epithelium may be potentially altered in keratoconus [8].
Meibomian gland dysfunction (MGD) is the leading cause of evaporative dry eye. It is characterized by terminal MG duct obstruction and/or altered changes in the meibum secretions, leading to tear lm instability, thus having the potential of initiating and perpetuating the vicious circle of dry eye disease, which may ultimately lead to in ammation and ocular surface damage [9][10][11][12].
Our purpose in this study was to investigate the MG function and morphology in keratoconic eyes and evaluate any possible in uence of MGD in the ocular surface damage and tear lm instability that may be encountered in these eyes.

Patient Selection
This study was approved by the Institutional Review Board Committee of the Ankara University and was performed according to the ethical principles of the Declaration of Helsinki. The medical records of consecutive keratoconus patients who were examined using a standardized protocol between the years 2016 to 2020 at the Cornea, Contact Lens and Refractive Surgery Service of Ankara University School of Medicine were retrospectively reviewed. Exclusion criteria included history of dry eye diagnosis or ocular surgery, soft or rigid contact lens wear, use of systemic medications that may cause dry eye or MG dysfunction, presence of rosacea, allergic, atopic or vernal keratoconjunctivitis or any other coexisting ocular surface disease.
All patients underwent complete ophthalmologic evaluation including slit lamp examination (SLE, at 40× magni cation), corneal tomography (The Oculus Pentacam® Oculus Optikgeräte GmbH, Wetzlar, Germany) and noncontact meibography (Sirius Scheimp ug Camera; Bon Optic VertriebsgmbH, Lübeck, Germany). The diagnosis of keratoconus was made based on the presence of slit lamp biomicroscopy ndings such as Fleischer ring, stromal thinning or Vogt striae, together with corneal tomographic evidence of ectasia. If both eyes of one patient had keratoconus, only the eye with more advanced keratoconus was included into the study. Keratoconic eyes were staged according to the Amsler-Krumeich classi cation [13].

Slit lamp biomicroscopy Findings and Functional Analysis
Patients were asked to rate their dry eye symptoms using the Ocular Surface Disease Index [OSDI] questionnaire. Ocular examinations were performed in the following order; evaluation and measurement of spectacle-corrected distance visual acuity (CDVA), non-invasive break up time (NIBUT) and MG morphology measurements using non-contact meibography (Sirius Scheimp ug Camera; Bon Optic VertriebsgmbH, Lübeck, Germany), Scheimp ug corneal tomography (Pentacam® Oculus Optikgeräte GmbH, Wetzlar, Germany), slit-lamp biomicroscopy, uorescein break up time (TBUT), ocular surface staining, and Schirmer I test. Corneal and conjunctival staining was graded according to the Oxford grading scale [14].

Morphological Analysis
The best recorded images of the upper and lower lids were chosen and the morphological characteristics of the meibography images were evaluated by one examiner (GO) who was blinded to the medical records of the patients. All morphological assessments were obtained from the middle two thirds of the eyelids on meibography.
The "total meiboscore" was the percentage of partial or total loss of MG on a de ned area on the everted eyelid. Degree of meibomian gland loss was calculated with the software of the device automatically and given as percentage. The meiboscore was graded as grade 0 (no loss of MG), grade 1 (the affected area, less than 33% of the total area occupied by the MG), grade 2 (the affected area between 33% and 66% of the total area occupied by the MG), and grade 3 (the affected area more than 66% of the total area occupied by the MG) [15].
"Partial gland" was de ned as partial loss of meibomian gland from the ori ce or fornix, and "gland dropout" was de ned as complete loss of meibomian gland from ori ce to fornix. Scores were identi ed by the number of affected MGs according to grading suggested by Arita et al. 15 Partial loss of MGs was graded as grade 0 (no partial glands), grade 1 (less than three partial glands), grade 2 (three or more partial glands and less than three partial glands with loss of half or more of the full length), or grade 3 (three or more partial glands with loss of half or more of the full length) ( Figure 1).
Gland dropout was graded as grade 0 (no gland dropout), grade 1 (less than three gland dropouts), or grade 2 (three or more gland dropouts) ( Figure 2). Glands that appeared curly compared to other glands of the patient's eye were de ned as "curled glands" and glands that appeared at least twice the thickness of other glands of the same patient eye were de ned as "gland thickening" and scores were identi ed by the number of affected MGs according to grading of Aslantürk and Uçakhan; grade 0 (no affected gland), grade 1 (less than three curled glands or thickened glands), or grade 2 (three or more curled glands or

Statistical Analysis
Descriptive statistics were expressed as mean ± standard deviation and median (min-max) for continuous variables, and as number and percentage for categorical variables. Statistical tests and correlation coe cients were chosen according to whether the data distribution was normal or not. Categorical variables were compared using Chi-Square or Fisher's Exact Test as appropriate. The differences between the groups were evaluated by Mann-Whitney U test. P-values less than 0.05 were considered as statistically signi cant. Statistical analyses were performed using SPSS (Statistical Package for Social Sciences; version 11.5).

Results
A hundred eyes of 100 patients with the diagnosis of keratoconus (55 males, 45 females) and 100 eyes of 100 healthy control subjects (55 males, 45 females) were included in this retrospective study. The mean age of the patients was 22.6 ± 6.0 years (range; 11 to 43 years) in the keratoconus group and 23.5 ± 6.9 years (range; 11 to 45 years) in the control group. There were no statistically signi cant differences between the two groups in terms of age and sex (p > 0.05).

Slit Lamp Biomicroscopy Findings And Functional Analysis
The functional parameters analyzed in the keratoconus and control groups are summarized in Table 1. Accordingly, the mean OSDI scores, NIBUT, TBUT, and corneal staining were signi cantly worse in the keratoconus group (p < 0.05) compared to healthy eyes. The non-contact meibography ndings were evaluated separately for the upper and lower eyelids. In the upper eyelids, MG loss was 24.0 ± 15.7% in the keratoconus group and 13.7 ± 5.9% in the control group (p < 0.001). In the lower eyelids, MG loss was 34.0 ± 16.2% in keratoconus group and was 20.4 ± 11.2% in the control group (p < 0.001) (Fig. 5). Table 2 shows the mean meiboscore, partial gland, gland dropout, as well as curled and thickened gland scores in the two groups. The mean meiboscore, partial gland, gland dropout, and thickened gland scores of the upper/lower eyelids were signi cantly higher in keratoconus group compared to the control group (p < 0.05). The mean curled gland score of the upper/lower eyelids was also higher in the keratoconus group compared to the control group, however, this difference reached statistical signi cance only for the lower eyelids.  Table 3 shows the correlations between the functional and morphological parameters studied in the keratoconus group. Among functional parameters, NIBUT correlated with partial or total MG loss, and meiboscore for both the upper and lower eyelids (p < 0.05), whereas, TBUT correlated with meiboscore and partial gland loss only in the upper lid (p < 0.05).

Correlations Between Functional And Morphological Alterations And Keratoconus Severity
Of the total 100 eyes, 30% had grade 1, 25% had grade 2, 25% had grade 3, 20% had grade 4 keratoconus. Table 4 shows the correlations between functional and morphological alterations in the upper and lower eyelids and the keratoconus severity. Correspondingly, there were statistically signi cant correlations between disease severity and OSDI scores, TBUT, NIBUT, meiboscore, partial gland, and gland thickening scores for upper/lower eyelids and gland dropout scores for lower eyelids (p < 0.05).

Discussion
In our study, there were signi cant functional and morphological alterations in the MG of keratoconic eyes compared to normals. The keratoconus group showed signi cantly higher OSDI scores and lower NIBUT, TBUT, greater corneal and conjunctival staining compared to the control group. Meiboscore, partial gland, gland thickening, and gland dropout scores were also signi cantly higher in the keratoconus group compared to age-matched controls both for the upper and lower eyelids. Furthermore, OSDI scores, tear lm instability and most morphological alterations correlated with disease severity.
Keratoconus is a multifactorial disease with genetic, environmental, in ammatory, hormonal risk factors and eye rubbing playing role in etiopathogenesis [17]. In the literature, few studies reported on increased dry eye symptoms [7,18], tear lm instability [6,19], ocular surface staining [6, 7] and conjunctival metaplasia with goblet cell loss in keratoconus patients [5][6][7]. indicating involvement of whole ocular surface epithelium and tear lm in the disease process in these eyes. Although the reason for such ndings is not clear, in ammatory changes were detected in the tear lm of keratoconic eyes, which were suggested to lead to goblet cell loss and alterations in mucin secretion with resultant tear lm instability [3,4]. On the other hand, whereas, MGD is considered the most common reason for tear lm instability, and evaporative dry eye, knowledge on the role of MGD in signs and symptoms of ocular surface disease in keratoconus is scarce [20].
Meibomian gland disease is a chronic condition of the MG, characterized by terminal duct obstruction and/or qualitative or quantitative changes in glandular secretion. Although traditionally MGD is diagnosed on the basis of subjective symptoms and slit lamp biomicroscopy ndings such as; lid margin abnormalities, the condition of the gland ori ces, and meibum grade, rapid objective evaluation of the MG with high reproducibility is only possible with noncontact meibography [21]. This method is now widely adopted in clinical practice for evaluation of meibomian gland-related diseases. Meibomian gland dropout on meibography has been reported to be correlated with NIBUT, TBUT, lipid layer thickness, meibum expressibility and tear lm osmolarity [22][23][24].
Previously, Mostafa et al reported increased OSDI scores, decreased TBUT and Schirmer test results, and increased MG dropout in keratoconus patients compared to controls. In their study, in which only upper eyelid meibography measurements were performed, meiboscore grades were not signi cantly different between the two groups. The authors could not demonstrate any correlation between TBUT, Schirmer test results, gland dropout or gland distortion versus severity of keratoconus. In our study, we observed signi cant differences in OSDI, TBUT, and MG morphology between keratoconic eyes and normals. Among the MG morphologic parameters meiboscore and partial gland, gland dropout and, gland thickening scores were signi cantly higher in KC group compared to controls in both upper and lower eyelids. Furthermore, we evaluated NIBUT as has been suggested by the TFOS DEWS II report [26] to evaluate tear lm stability, and the NIBUT measurements in these patients showed moderate to strong correlations with upper lid and lower lid meiboscore and partial gland scores. We could also demonstrate signi cant correlations between meiboscore, partial gland and gland thickening scores of both the upper and lower eyelids and the severity of keratoconus, with worsening of meibography ndings as the disease stage increased.
To our knowledge, our study is the rst study showing such correlations between OSDI, NIBUT, meibography ndings and keratoconus severity. The reason for the morphological changes in MG is not clear, however, ocular surface in ammation initiated by tear lm instability, or directs mechanical irritation of the lids by the conic cornea or vigorous eye rubbing may be possible underlying mechanisms. Although tear lm dysfunction or MGD cannot cause keratoconus, such ocular surface problems may further exacerbate the disease through increased eye rubbing.
One of the limitations of the present study is its retrospective nature. We could not reach lid margin abnormality data of all patients from patient charts and therefore did not include slit lamp biomicroscopy gradings of MG ori ces or secretion quality into the study, which could have increased the accountability of our current ndings. However, our patients were not MGD patients per se, furthermore, at the outpatient clinic we found evaluation of gland ori ces via observation of meibography images much easier and reliable as compared to evaluation at the slit lamp with or without digital pressure on the lids. Therefore, we believe, the outcomes of our study con rm the presence of signs and symptoms of tear lm instability in keratoconus patients and add new information to our knowledge of MG morphology in these eyes as measured by non-invasive meibography.
In summary, keratoconus seems to be associated with signi cant morphological alterations in MG of both upper and lower eyelids. Furthermore, the morphological alterations in these eyes seemed to correlate with symptoms, tear lm instability (particularly NIBUT) and keratoconus stage. Although further longitudinal studies are required to elucidate the mechanism underlying the association between keratoconus and MGD, the present ndings indicate the possibility that MGD may be one of the underlying factors in dry eye symptoms and tear lm instability in patients with keratoconus. Representative images of partial gland loss grading marked with black stars. (A,E) Partial gland loss grade 0, no partial glands. (B,F) Partial gland loss grade 1, less than three partial glands. (C,G) Partial gland loss grade 2, three or more partial glands and less than three partial glands with loss of half or more of the full length. (D,H) Partial gland loss grade 3, three or more partial glands with loss of half or more of the full length   Comparisons of the mean MG loss for the upper (blue) and lower (green) eyelids in the control group and different stages of keratoconus.