Purpose To study the relationship between epithelial thickness (ET) and corneal thickness (CT) in healthy fellow eyes of patients with unilateral bullous keratopathy (BKP) and healthy subjects.
Methods Seventeen BKP patients (nine males, eight females, 73.2±10.4 years) and 40 healthy individuals (20 males, 20 females, 69.5±9.8 years) were included. All participants received anterior segment optical coherence tomography and specular microscopy with calculation of endothelial cell density. ET, CT, stromal thickness and ET-to-CT ratio were defined automatically (within 2 mm central area). Central epithelial thickness (CET) and central corneal thickness (CCT) were measured manually at the corneal centre and stromal thickness and CET-to-CCT ratio were calculated.
Results In healthy fellow eyes of BKP patients compared with eyes of healthy individuals ET was statistically significantly lower (p<0.001) while CT was statistically significantly higher (by 28.9 and 30.9 µm in 2 mm zone and corneal centre, respectively). Both 2 mm ET-to-CT ratio (0.091±0.01 and 0.10±0.004, respectively [p<0.001]) and CET-to-CCT ratio (0.083±0.006 and 0.97±0.005, respectively [p<0.0001]) were statistically significantly lower in healthy fellow eyes of BKP patients compared with eyes of healthy individuals. To identify healthy fellow eyes of BKP patients, area under curve for CET-to-CCT ratio and CET was 0.94 and 0.80, respectively (p=0.01), and for 2 mm ET-to-CT ratio and 2 mm ET was 0.91 and 0.80, respectively (p=0.03).
Conclusion Decreased СET-to-СCT ratio resulting from simultaneous epithelial thinning and stromal thickening without significant corneal thickening indicates subclinical dysfunction of corneal endothelium in healthy fellow eyes in unilateral BKP patients.
- anterior segment optical coherence tomography
- bullous keratopathy
- corneal epithelium thickness
- corneal endothelium
- corneal thickness
- fuchs endothelial dystrophy
Statistics from Altmetric.com
- anterior segment optical coherence tomography
- bullous keratopathy
- corneal epithelium thickness
- corneal endothelium
- corneal thickness
- fuchs endothelial dystrophy
Corneal epithelium considered to be the most plastic corneal layer and a sensitive indicator for a number of corneal and anterior segment abnormalities, for example, dry eye and keratoconus.1 2 However, clinical application of corneal epithelium measures faces difficulties due to the broad range of the normal value. Although several predictors for central epithelial thickness (CET) have been described previously, only central corneal thickness (CCT) which demonstrates statistically significant correlation with CET,3 is strong enough to be used in clinical practice. Taking into account such a correlation, for improvement of the epithelial metric, it appears to be reasonable to use CET-to-CCT ratio rather than absolute CET value.
One of the conditions associated with profound changes in corneal morphometric parameters including corneal thickness is bullous keratopathy (BKP).4 There are several possible causes for endothelial decompensation leading to BKP, including, surgical trauma, Fuchs endothelial dystrophy and anterior segment inflammation.5 Unilateral BKP can be promoted by cataract surgery in patients with preoperatively decreased endothelial cell density or Fuchs endothelial dystrophy (and can be spontaneous in this condition).6 In such case, subclinical changes (at least loss of endothelial cells) in healthy eye nevertheless can be found.7 Therefore, patients with unilateral BKP represent an adequate model for evaluating subclinical changes of corneal morphology associated with endothelial cells loss. We believe that this data can not only facilitate early diagnosis of Fuchs endothelial dystrophy but also allow for identification of subclinical decompensation of endothelial cells in general and possibly improve our understanding of the clinical course of such conditions.
In this study using spectral domain optical coherence tomography (OCT) we evaluated the ratio of the epithelial to corneal thickness as well as epithelial, stromal and total corneal thickness in healthy fellow eyes of patients with unilateral BKP and compared with those in healthy eyes of age-matched individuals. We also evaluated the relationships between OCT parameters and loss of endothelial cells and, finally, the ability of epithelial to corneal thickness ratio to identify eyes with subclinical endothelial dysfunction.
Material and methods
In this prospective cohort study, we included patients with unilateral BKP and healthy age-matched individuals. Exclusion criteria for healthy fellow eyes of patients with unilateral BKP were as follows: any biomicroscopically detected corneal abnormalities (with exception for alteration of corneal endothelium, precipitates and pigmentary granules on posterior corneal surface), contact lens wearing, any known diseases or conditions of anterior eye segment (excluding cataract), specifically any inflammatory or allergic diseases of anterior eye surface, history of previous refractive corneal or ocular adnexa surgery, abnormalities of eyelids (for example blepharitis or its symptoms) and any known topical medication. In the control group we included age-matched healthy individuals who also did not meet the following exclusion criteria: (1) any known or suspected dry eye disease (including reported experience of dryness, stinging or burning in any eye), (2) endothelial cell density of 2000 cells/mm2 or lower and (3) bilateral pseudophakia. Only phakic eyes of healthy individuals were included in the study. The study adhered to the tenets of the Declaration of Helsinki and was approved by the Local Ethics Committee.
Anterior segment optical coherence tomography
All participants received a comprehensive ophthalmic examination, anterior segment OCT and specular microscopy with calculation of endothelial cell density. Anterior segment OCT was performed for both eyes using widefield pachymetry protocol (8 radial averaged 9 mm B-scans) with Avanti RTVue XR (Optovue, Fremont, California). Endothelial cell density for the central corneal area was calculated automatically with specular microscope Perseus (CSO, Firenze, Italy) as a mean of three consecutive image acquisitions.
Outcome measures in this study were: (1) automatically calculated epithelial thickness (ET) within 2 mm central area (2 mm ET), (2) automatically calculated corneal thickness (CT) within 2 mm central area (2 mm CT), (3) manually measured CET and (4) manually measured CCT. CET and CCT were measured in masked fashion using built-in calliper tool on one B-scan of widefield pachymetry protocol in three points within 1 mm area as was described earlier.3 Stromal thickness was defined as the difference between corneal and epithelial thickness based on both automatically provided and manually (central stromal thickness) determined measures.
Epithelial to corneal thickness ratio was calculated using: (1) automatically generated values as the ratio of 2 mm ET to 2 mm CT (2 mm ET-to-CT ratio) and (2) using manually defined values as CET-to-CCT ratio.
Statistical analysis was performed using MedCalc V.18.4.1 (MedCalc Software). All data presented as a mean±SD. Kolmogorov-Smirnov test was used to confirm the normality of distribution of the data. Independent samples t-test was used to compare continuous variables between study groups. For binary variables, the Fisher’s exact test was performed.
Spearman correlation coefficient was calculated between epithelial and corneal thickness in eyes of healthy individuals and in healthy eyes of BKP patient. Statistical power calculation has indicated that the sample size of 37 or more eyes of healthy individuals needed to find a correlation coefficient of 0.55 or higher with a power of 0.95 at a significance level of 0.05.
Additionally, we calculated correlation coefficient between epithelial thickness, corneal thickness as well as their ratio and endothelial cell density in the entire group of all eyes included in the study. Receiver operating characteristics (ROC) curves were constructed to evaluate the diagnostic accuracy of CET-to-CCT ratio, 2 mm ET-to-CT ratio and CET in the differentiation of healthy fellow eyes of BPK patients from eyes of healthy individuals. The difference at p value of 0.05 or lower was defined as statistically significant.
Seventeen patients with unilateral BKP (nine males and eight females, mean age 73.2±10.4 years) and 40 healthy individuals (20 males and 20 females, mean age 69.5±9.8 years) were included in this study. No statistically significant differences in age or male-to-female ratio were found between the groups (table 1). In unilateral BKP group, all healthy fellow eyes and 11 out of 17 BKP eyes were phakic, while among healthy individuals 28 eyes were pseudophakic (p<0.001).
The endothelial cell density in healthy fellow eyes of BKP patients was statistically significantly lower than in healthy individuals (1582±587 cells/mm2 and 2531±252 cells/mm2, respectively, p<0.001). There was no statistically significant difference in the endothelial cell density between phakic and pseudophakic eyes of healthy individuals. Both 2 mm CT and CCT demonstrated statistically significant differences between the groups (p=0.001 and p<0.001, respectively). In healthy fellow eyes of BKP patients compared with eyes of healthy individuals, stromal thickness (both automatically and manually determined) was statistically significantly higher (p<0.001) while epithelial thickness (both automatically determined 2 mm ET and manually measured CET) was statistically significantly lower (p<0.001, respectively) (figure 1). There was no statistically significant difference in 2 mm ET, 2 mm CT, CET and CCT between phakic and pseudophakic eyes of healthy individuals (p>0.05).
In the group of healthy individuals, we found a moderate positive statistically significant correlation between 2 mm ET and 2 mm CT (r=0.55, p<0.001) and CET and CCT (r=0.59, p<0.001), while in healthy fellow eyes of BKP patients such correlation was not found (r=0.18, p=0.47).
In healthy fellow eyes of BKP patients compared with eyes of healthy individuals both 2 mm ET-to-CT ratio (0.091±0.01 and 0.10±0.004, respectively [p<0.001]) and CET-to-CCT ratio (0.083±0.006 and 0.97±0.005, respectively [p<0.0001]) were statistically significantly lower (figure 2).
ROC analysis showed that in the diagnosis of subclinical dysfunction of corneal endothelium (differentiation of healthy fellow eyes of BKP patients from eyes of healthy individuals) epithelial to corneal thickness ratio has a greater area under curve (AUC) compared with both CET and 2 mm ET. Pairwise comparison of ROC curves showed statistically significant difference in AUC for CET-to-CCT against CET (0.94 and 0.80, respectively, p=0.01), CET-to-CCT ratio against 2 mm ET (0.94 and 0.80, respectively, p=0.01) and 2 mm ET-to-CT ratio against 2 mm ET (0.91 and 0.80, respectively, p=0.03) (figure 3). There was statistically significant difference in AUC for CET-to-CCT ratio against CST (0.94 and 0.76, respectively, p=0.007) and CET-to-CCT ratio against 2 mm ST (0.94 and 0.76, respectively, p=0.01). ET-to-CT ratio within 2 mm central zone also demonstrated statistically significantly larger AUC against CST (0.91 and 0.76, respectively, p=0.03) and 2 mm ST (0.91 and 0.76, respectively, p=0.03) (figure 3).
There was statistically significant negative correlation between 2 mm CT (r=-0.59, p<0.001) and corneal endothelial cell density and between CCT (r=-0.58, p<0.001) and corneal endothelial cell density. Statistically significant positive correlation was found between 2 mm ET-to-CT ratio and corneal endothelial cell density (r=0.73, p<0.001) and between CET-to-CCT and corneal endothelial cell density (r=0.67, p<0.001). Correlation between epithelial thickness and corneal endothelial cell density was not statistically significant for both 2 mm ET (r=0.44, p=0.01) and for CET (r=0.34, p=0.05) (figure 4).
In this study, the CET-to-CCT ratio has proved to be highly accurate in distinguishing healthy fellow eyes of patients with unilateral BKP from eyes of healthy individuals and, therefore, this ratio can be used as a diagnostic criterion for subclinical dysfunction of corneal endothelium. Healthy fellow eyes of patients with unilateral BKP, which have close to normal corneal thickness, demonstrated increased thickness of the corneal stroma and thinner epithelium when compared with healthy age-matched controls. However, each of these indices taken alone is only moderately accurate in prediction of endothelial cell density. Therefore, maintaining a stable CET-to-CCT ratio can be adopted as a function of endothelial cell density.
The increase of corneal thickness in endothelial decompensation is a well-known phenomenon which contributes substantially to the clinical appearance of BKP.4 Since corneal endothelium has a barrier role, its loss consequently leads to stromal hydration, oedema and thickening. Although the loss of endothelial cells below a certain threshold is needed for the manifestation of BKP, one may suggest that functional competency of the corneal endothelium as a whole is a continuous characteristic which is depending on the number of endothelial cells.8
The link between corneal thickness and loss of endothelial cells in eyes with Fuchs endothelial dystrophy without BKP, but with substantially increased corneal thickness, has been already described.9 In our study, this was also confirmed for eyes with generally normal corneal thickness. Moreover, the epithelial to corneal thickness ratio appears to have a stronger correlation with endothelial cell density compared with corneal thickness alone. This is not surprising since the ratio includes two parameters both of which are changing simultaneously.
The decrease of epithelial thickness can be explained by dysfunction of corneal nerve plexus in subclinical stromal oedema which results in neurotrophic keratopathy and epitheliopathy. This is particularly interesting in the context of the study of Bucher et al, who demonstrated substantial changes in the density and anatomy of the subbasal corneal nerve plexus in advanced Fuchs endothelial corneal dystrophy.10
Interestingly, in this study, a correlation between CET and CCT in healthy aged individuals (mean age 67.4±20.7 years) was found to be at least not lower than that in young adults (mean age 33.9±12.1 years) in the previous study, r=0.55 and r=0.53, respectively. Previously, the lack of data about the correlation between CET and CCT in aged population has limited the application of CET-to-CCT ratio in such patients.3 Indeed, if the correlation were weak, diagnostic application of CET-to-CCT ratio in elderly population would be inaccurate.
The fact that both 2 mm ET and CET are statistically significantly lower in healthy fellow eyes of BKP patients allows us to consider epithelial thinning as a diagnostic criterion for subclinical dysfunction of the corneal endothelium. However, epithelium to corneal thickness ratio in general and especially CET-to-CCT ratio demonstrated better accuracy in differentiating BKP patients from healthy individuals. It is also worth noting that although manual measurements make the analysis grader-dependent, this enables us to minimise the influence on analysis of automatic segmentation errors and artefacts associated with tear fluid and central corneal reflection.3
We suggest that the CET-to-CCT ratio may become a relatively simple and reliable diagnostic indicator for endothelial dysfunction, for example, for subclinical Fuchs dystrophy. At this moment, to the best of our knowledge, endothelial dysfunction is the only known reason for the decrease in the CET-to-CCT ratio. Interestingly, the ratio of epithelial to corneal thickness have shown a correlation with the density of corneal endothelial cells. This could be expected since hydration of the cornea, which at least partially determines its thickness, depends on the functional competency of the corneal endothelium. We suggest that epithelial to corneal thickness ratio indicates not only the endothelial cell density but also the functional state of corneal endothelium and therefore may be considered as an indicator for further loss of endothelial cells since at cellular level, cell death is typically associated with previous functional decompensation.
This study has several limitations. First, even though the mean endothelial cell density in healthy fellow eyes of BKP patients was substantially below the normal limits, we cannot exclude the possibility that in some patients with relatively high endothelial cell density endothelium was not compromised. This could be true for pseudophakic patients, for example, in whom BKP might have been caused only by endothelium damage during cataract surgery. In our study, we did not limit inclusion of patients with BKP by endothelial cell density. We also did not exclude patients with pseudophakia, since patients with Fuchs dystrophy could be among them. Nevertheless, we believe that the inclusion of BKP patients with relatively high endothelial cell density (but less than normal) in the healthy fellow eye does not affect the main conclusions of this study. Second, study groups were not matched in accordance to the phakic status of the eyes included. This fact does not influence the results of the study, since only the group of healthy individuals included pseudophakic eyes which simultaneously demonstrated high endothelial cell density. Moreover, there was no statistically significant difference in endothelial cell density between phakic and pseudophakic eyes of healthy individuals. Third, it is not known if the changes in epithelial and corneal thickness and consequently in their ratio appear in patients without unilateral BKP and, if they do, how early they appear before BKP manifestation. Fourth, it is worth bearing in mind that epithelium thickness is strongly dependent on tear film stability and eyelid abnormalities. Since dry eye and blepharitis are widely distributed among the elderly population, in routine practice personalisation of epithelial metric can be limited. Therefore, patients with anterior eye abnormalities were excluded from this study.
This study also has strengths, namely, that we performed analysis based on both manually measured and automatically generated values. The results obtained with manual and automatic measures were very close to each other, and this justified the manual performance in the evaluation of the CET-to-CCT ratio.
In conclusion, in this study, we found that anatomical changes of the cornea associated with loss of corneal endothelial cells appear before clinical manifestation of BKP. These changes include not only stromal thickening but also epithelial thinning, without significant increase of corneal thickness. Taken together as the CET-to-CCT ratio, these parameters provide an accurate gradual indicator for subclinical dysfunction of the corneal endothelium.
Contributors DSM, ANK and EVK designed the study. DSM and MAB collected the data. DSM and MAB analysed the data. DSM and AVK drafted the manuscript. DSM, ANK and EVK critically reviewed and approved the manuscript.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Obtained.
Ethics approval The Ethics Committee of Military Medical Academy.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Data are available upon request.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.