Surgical Success Evaluation in Limbal Stem Cell Deciency Due to Chemical Injury – How to Be Accurate?

Purpose: To evaluate limbal stem cell transplantation (LSCT) success in limbal stem cell deciency (LSCD) due to chemical injury in a tertiary eye care center in Turkey by using ‘Limbal Stem Cell Working Group’ LSCD grading system. Methods: Medical records of 80 eyes of 80 patients who underwent LSCT for LSCD secondary to chemical injury were included. The patients were divided into 3 groups according to performed surgery as limbal autograft, limbal allograft, cultivated limbal epithelial cell transplantation (CLET). Surgical success was dened as improvement in the post-operative 1 st year LSCD stage. Results: The mean age of the patients was 37.9±15.7(4-71) with a Male/Female ratio of 2.4. Forty-ve(56.3%) patients were injured with alkaline and 16(20%) were injured with acid substance. The mean follow-up time was 60.3±30.6(6-118.6) months. Limbal autograft, allograft and CLET were performed in 58(72.5%), 12(15%) and 10(12.5%) eyes, respectively. The interval between injury and surgery in limbal autograft, limbal allograft and CLET were 43.3±94.1(0.5-592); 14.5±10.6(2.4-32.5) and 122.8±158.9(21.1-504) months, respectively (p=0.02). The overall surgical success rate was 65%. Surgical success rates in each groups were 65.5%, 41.7% and 90%, respectively (p=0.03). Conclusion: It is vital to accurately determine the stage of the LSCD in order to evaluate the surgical success rates. Surgery type and a longer time interval between the chemical injury and surgery seems to be the most important factors associated with a higher surgical success rate. Despite the limited subject number for subgroups, the results were remarkable to emphasize the signicance of the novel LSCD grading system.


Introduction
Chemical injury of the ocular surface is a true ocular emergency that requires urgent and intensive management. [1] Dua classi cation is used for grading the severity of ocular chemical injury and also may predict the visual prognosis. [2,3] Severe ocular chemical injury may cause limbal stem cell de ciency (LSCD), which occurs due to dysfunction of limbal stem cells and impairment of the limbal barrier. LSCD secondary to chemical injury is usually unilateral, and may either be total or partial. [4][5][6] The accurate evaluation of LSCD has a great importance in choosing the proper management of the disease. Although medical treatment may be useful in mild LSCD, replacement of the limbal stem cells via graft or cell transplantation is mandatory in severe LSCD. [7][8][9][10][11] There was not a consensus about the grading of LSCD until 2019. By then, to evaluate the severity of LSCD, a new grading system was de ned by "Limbal Stem Cell Working Group". [9] After limbal stem cell transplantation (LSCT), the overall surgical success rates vary between 50-80% changing according to the tissue source. Beside the tissue source selected for LSCT, there are several other factors such as, surgery type, LSCD grade, the duration between the chemical injury and surgical intervention, that may affect the success rates as well. [10][11][12] Also for accurate evaluation of the surgical success, objective criteria use is particularly vital. [13,14] The aim of this study is to accurately evaluate the LSCT success in LSCD due to chemical injury in a tertiary eye care center in Turkey by using 'Limbal Stem Cell Working Group' LSCD grading system.

Materials And Methods
This retrospective study was conducted with the approval of the Ethics Committee of Ege University and in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants.
Patients who were admitted to Ege University Department of Ophthalmology between the years of 2007 and 2019 with LSCD secondary to ocular chemical injury and underwent LSCT were included. Medical records of patients including age, gender, nature of the chemical agent, accompanying ocular injuries (eyelid malformation, symblepharon, etc.), the time interval between injury and LSCT, and transplantation methods were reviewed. Pre-and post-operative (6th month, 1st year, 3rd year follow ups) best-corrected visual acuity (BCVA) measurements (Snellen chart) as well as intraocular pressure (IOP) measurements were documented. Anterior segment photographs of the patients were analyzed by two different researchers, and chemical injury grades and LSCD stages were recorded. Patients with Grade 3 and higher chemical injury grade according to Dua classi cation [2] were included. Pre-and post-operative (6th month, 1st year, and 3rd year) LSCD stages were categorized according to the staging system established by the 'Limbal Stem Cell Working Group'. [9] According to this classi cation, Stage 1 was de ned as normal corneal epithelium within the central 5 mm zone of the cornea. Additionally, according to the degree of limbal involvement this stage was divided into three subgroups as A, B and C. In Stage 2, the central 5 mm zone of the cornea and also limbus are both affected. In Stage 3, the entire corneal surface is affected. In the presented study, medical records of patients with Stage 1B and/or higher LSCD stages at pre-operative examination were included.
The performed LSCT treatments were limbal autograft, living relative limbal allograft and cultivated limbal epithelial transplantation (CLET). CLET was performed as an explant culture, which was expanded on human amniotic membrane with autologous serum.
'Successful surgical outcome' was de ned as improvement in the post-operative 1st year LSCD stage, regardless of visual acuity improvement. The patients were divided into 3 groups according to the used LSCT type as Group 1: limbal autograft, Group 2: limbal allograft, and Group 3: CLET.
LSCD secondary to other causes rather than ocular chemical injury, follow-up less than 6 months, previous ocular surgery or other trauma history, and absence of anterior segment photographs or regular medical records were the exclusion criteria.
Descriptive statistics were given as mean, standard deviation, median, minimum, maximum, frequency and percentage. The Shapiro-Wilk test was used to test the normality assumption of quantitative variables. In univariate analyzes, one-way ANOVA was used for normally distributed variables, and Kruskal-Wallis test (Dunn test for paired comparisons) for non-normally distributed variables. Categorical variables were compared using the Pearson Chi-square test and Fisher's exact test. To examine relationship between quantitative variables, Spearman correlation analysis was performed. Univariate logistic regression analysis was performed to determine the factors affecting the surgical success rate. Parameters with p < 0.10 in univariate analyzes were included in the multiple logistic regression analysis.
IBM SPSS Statistics 25.0 (IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY: IBM Corp.) was used in the statistical analysis. P < 0.05 was considered signi cant.

Results
The mean age of the patients was 37.9 ± 15.7 (4-71) with a male to female ratio of 2.4. Forty-ve (56.3%) patients were injured with an alkaline and 16 (20%) patients were injured with an acid substance.
Overall, 52 of 80 patients (65%) showed surgical success according to the novel LSCD grading system described by 'Limbal Stem Cell Working Group'. The mean improvement in LSCD stage in patients with successful surgical outcome was 1.9 ± 0.9 (1-4). The LSCD stages were stable in 22 patients (27.5%) and decreased in 6 patients (7.5%). In these 6 patients with worsening LSCD scores, the mean decrease in the LSCD stage was 1.3 ± 0.8 (1-3). There was not any signi cant relationship between gender, age, nature of chemical agent, presence of symblepharon and chemical injury grade with surgical success (p > 0.05 for all categories). Preoperative LSCD stage did not have statistically signi cant effect on surgical success rates (p = 0.1). In a multivariate logistic regression analysis controlling for age, nature of chemical agent, presence of symblepharon and chemical injury grade; time interval between injury and transplantation predicted surgical success. Longer time interval between injury and transplantation, was associated with higher surgical success rates (p = 0.05).
The patients were divided into three groups according to the type of surgery performed. Limbal autograft was performed in 58 (72.5%) eyes, limbal allograft was performed in 12 (15%) eyes, and CLET was performed in 10 (12.5%) eyes (Fig. 1). There was no signi cant difference between these three groups in terms of mean age, gender, nature of chemical agent, chemical injury grades and pre-operative LSCD stages (p > 0.05 for all categories). The time interval between injury and surgery in limbal autograft, limbal allograft and CLET were 43.3 ± 94.1 (0.5-592); 14.5 ± 10.6 (2.4-32.5); 122.8 ± 158.9 (21.1-504) months, respectively; being signi cantly higher in CLET group compared to other 2 groups (p = 0.02). (Table 1)  ). In 1 patient (9.1%), BCVA and LSCD stage remained the same. No increase in LSCD stage and no decrease in BCVA were observed after PK in any patients. The increase in BCVA after PK was found to be signi cantly higher in those with a higher initial LSCD stage (p = 0.04).

Discussion
Chemical injury of the ocular surface is a signi cant public health problem with a serious economic effect due to prolonged hospitalization, long-term medical treatment, multiple surgery requirement, workday loss, and even blindness. [15] In severe ocular chemical injury, the main cause of blindness is LSCD. The current and de nitive treatment of LSCD is the replacement of healthy limbal stem cells. [7,16,17] Young male individuals are at a higher risk of ocular chemical injuries as well as other traumas, especially due to work accidents. [18,19] In accordance with the literature, in the present study the majority of the patients (n = 57,71.2%) were male and the mean age of the patients was 37.9 ± 15.7. [20] In chemical ocular surface injury, exposure to alkaline agents is reported to be more frequent than acid agents.[6, 21] In the present study, 56.3% of the overall injuries were caused by alkaline substances in consistent with the literature.
The severity of the injury generally depends on exposure time and type of chemical agent. [1] Potent corrosive agents (mostly alkaline substances) may penetrate deeper in the eye. Therefore, it develops prolonged in ammation not only on the ocular surface, but also in deeper structures causing more severe LSCD and other complications such as glaucoma and cataract formation. [22,23] In the present study, the eyes with higher chemical injury grades tend to have higher LSCD stages. For evaluating the surgical success, different success criteria for LSCT were de ned among literature. While some authorities described the anatomical improvement in ocular surface as surgical success; in other studies, functional improvement and the patient-reported symptoms are used to evaluate surgical success. [13] In a recent meta-analysis including 40 clinical studies (2202 eyes), the overall success rate of all LSCT was found to be 67.4%, and improvement rate of ocular surface was found to be 74.5%. [14] Furthermore, the lack of standardized criteria for evaluation of LSCD made it di cult to compare the outcomes of different treatment modalities. Although the increase in visual acuity was considered as a success criterion in most previous studies, LSCT actually aims to treat LSCD ndings, not to improve the vision. Stromal opacities are the main cause of poor vision in patients with severe ocular chemical injury, so the absence of visual improvement does not actually indicate any treatment failure. [14,24,25] Recently in 2019, 'Limbal Stem Cell Working Group' established an objective scale to evaluate LSCD. [9] This staging system provided insight not only in evaluating surgical success, but also in the diagnosis and de nition of LSCD, and determination of the appropriate treatment option. [9] There are several surgical intervention options for LSCD. Limbal autograft is the most commonly performed surgical option in unilateral cases. [25][26][27] In consistence with the literature, limbal autograft was the most common surgical intervention in the present study.
Surgery type is one of the most important factors that affects success rate. Many different techniques have been described according to the source of the cells and the carrier tissue. [7] While limbal autograft is generally preferred in unilateral LSCD; in bilateral cases, limbal allograft is needed.
[28] The main disadvantages of limbal autograft are unrepeatability and LSCD risk at the contralateral healthy eye. [10] In limbal allograft, there are main problems such as risk of rejection and requirement of systemic immunosuppressive therapy use. [29] CLET, on the other hand, is a relatively new and promising method with a success rate of approximately 80%. [4] While repeatability is the most important advantage of CLET, the main disadvantage is the procedures' expensive nature. [30] In a meta-analysis including 40 clinical studies (2202 eyes), improvement of the ocular surface was 85.7% (33-100%) for limbal autograft, 57.8% (0-89) for limbal allograft and 84.7% (44-91) for CLET. [14] Herein, surgical success rates at the post-operative 1st year were; 65.5% for limbal autograft, 41.7% for limbal allograft and 90% for CLET. The improvement in postoperative LSCD stage was statistically higher in the CLET. Although there was no difference between three groups in respect to nature of the chemical agent, chemical injury grades and preoperative LSCD stages, the time interval between injury and limbal transplantation was signi cantly longer in CLET group. These results suggest that the successful surgical results of CLET may be associated with the longer time interval between chemical injury and LSCT surgery due to the in ammation regression by time, which provides the limbal graft survival better.
It is known that different factors may also affect LSCT outcomes, beside the surgery type. [31] Cheng et al. [32] reported that preoperative symblepharon grade and presence of in ammation both play a signi cant role in CLET outcomes. In their series with 80 eyes with symblepharon secondary to chemical or thermal injury; success rate of CLET was higher in eyes with grade 1/2 symblepharon than eyes with grade 3 and 4 symblepharon. [32] El-ho et al. [33] reported better nal BCVA results after limbal allograft in Dua-Grade 4 chemical injury than in Dua-Grade 5 chemical injury. In their study with 20 chemical injury patients who underwent limbal allograft, all patients that needed regrafting were Dua-Grade 5. They also reported that all eyes with delayed re-epithelization after limbal transplantation were alkali injury. [33] However, in the present study, there was no signi cant relationship between gender, age, nature of chemical agent (alkaline/acid/other), presence of symblepharon and chemical injury grade with postoperative 1st year surgical success rates.
It is known that the in ammation is the main cause of graft failure in LSCT. [34] And in the subacute phase of chemical injury, a low degree of in ammation may persist despite the use of anti-in ammatory therapy. [23] Therefore, most of the authorities have indicated that performing LSCT in the chronic period of chemical injury is much better for graft survival. [23,33] However, publications showing the direct effect of LSCT timing on LSCT outcomes are limited, and the ideal timing of LSCT after chemical injury has not been determined.
[26] Rao et al. [35] indicated that surgery at the acute phase of injury (< 4 months) is associated with delayed corneal re-epithelialization and poorer visual outcomes. Sejpal et al. [23] reported a high failure rate in eyes who underwent CLET within 4 months after chemical injury. On the contrary, Ozdemir et al.
[36] stated that early LSCT may prevent corneal neovascularization in chemical injury with large epithelial defects and small limbal ischemia. Herein, with 80 eyes, the longer interval between injury and transplantation was associated with the higher surgical success rates.
Also, the pre-operative LSCD stage may also be associated with the prognosis after LSCT. Due to the lack of LSCD grading consensus, there was not any detailed study that evaluates the pre-and post-operative LSCD grading association. In the present study, although lower LSCD stages at the initial examination was tent to have better surgical outcomes at 1st year, the statistical difference was not signi cant. To the best of our knowledge, this study will be one of the earlier studies, which accurately demonstrates the relationship between pre-operative LSCD stages and LSCT success by using the staging system established by 'Limbal Stem Cell Working Group'.
LSCT is an effective treatment for ocular chemical injury where only the corneal epithelium and limbus are affected. [37] It provides ocular integrity and improves visual acuity. However, if the corneal stroma is affected, PK is required for visual improvement. [24] PK simultaneously with LSCT is solely preferred in the presence of corneal perforation, since it has a high graft failure rate. [38,39] It is known that the presence of in ammation and vascularization in the recipient bed increases the risk of rejection of the corneal graft. [40] Therefore, it is recommended to perform PK once ocular surface stability is achieved with LSCT.
[38] Although there is no consensus about ideal timing of PK after LSCT, it is recommended to be performed at least 12 months later. [25] In addition, there is limited information in the literature regarding the potential effect of PK on limbal graft survival. Figuerdo et al. [37] recommended performing PK at least 12 months following CLET. They reported that performing PK at least 12 months after CLET did not negatively affect CLET survival and provided signi cant improvement in visual acuity. [37] In the present study, 11 of 80 patients underwent PK after LSCT, and the mean time between LSCT and PK was found to be 22.4 months. None of the patients had worsening of LSCD stage and BCVA after PK.
Although the number of patients who underwent PK was limited, it was enough to show that PK did not affect limbal graft survival. In 10 out of 11 patients, a signi cant increase in BCVA was achieved following PK. The increase in BCVA after PK was found to be higher in those with an advanced initial LSCD stage. In LSCD staging, involvement of the central 5 mm zone of the cornea was stated as the most important criteria. The LSCD stage is expected to be high in patients with central corneal involvement. So, in patients with effected central cornea, better visual acuity will be achieved after PK.
Although the main limitations of the study are its retrospective nature and the limited number of patients in the CLET and limbal allograft groups, the results of the current study are remarkable to highlight the impact of the novel LSCD grading system.
In conclusion, LSCT is the surgical treatment of severe ocular chemical injuries that cause LSCD. It is vital to determine the stage of the LSCD accurately in order to evaluate the surgical success rate. The most important factors affecting the outcomes of LSCT are, the method of transplantation and the time between surgery and injury. For this reason, in order to achieve higher success rates, rushing surgeries following chemical injury should be avoided.

Declarations
Funding: No funds, grants, or other support was received.

Con icts of interest/Competing interests:
The authors have no relevant nancial or non nancial interests to disclose.
Ethics approval: The study was approved by Institutional Ethics Committee of Ege University and followed the tenets of Helsinki Declaration Consent to participate: Informed consent was obtained from all individual participants included in the study.

Consent for publication:
All authors read the nal manuscript and give consent for the article to be published.