Article Text
PDF
Abstract
Background/aim To describe the clinical outcome of allogenic simple limbal epithelial transplantation (alloSLET) utilising tissue from cadaveric donor eyes after failed re-epithelialisation of the corneal surface.
Methods Medical records of 14 eyes from 14 patients treated for persistent corneal epithelial defects with alloSLET were reviewed. The primary outcome measure was complete epithelialisation of the corneal surface. Secondary outcome measures were best corrected visual acuity (BCVA) and postoperative side effects due to surgery or medical therapy.
Results Of the 14 eyes, 7 received alloSLET only and 7 alloSLET together with penetrating keratoplasty (PK). Thirteen (92.9%) of 14 eyes had an epithelialised corneal surface 3 and 6 months after surgery and 10 (71.4%) of 14 eyes displayed an epithelialised corneal surface 12 months after surgery. In both subgroups, alloSLET only and alloSLET with PK, respectively, 5 (71.4%) of 7 eyes had a stable corneal epithelium 12 months after surgery, respectively. Postoperatively, BCVA improved markedly in the whole patient collective. However, the increase was not significant when looking at the two individual subgroups. One patient lost his bandage contact lens several times within the first postoperative month and had a partial detachment of the amniotic membrane. The ocular surface of this patient failed to epithelialise. In three patients, limbal donor pieces translocated to the centre of the cornea, which possibly prolonged the improvement of BCVA.
Conclusion AlloSLET appears to be an effective treatment option in eyes with non-healing corneal epithelial defects when autologous limbal tissue is not available.
- ocular surface
- treatment surgery
- cornea
Statistics from Altmetric.com
Introduction
Limbal stem cell deficiency (LSCD) is a major clinical challenge in ophthalmology, especially when both eyes are involved.1 2 Regeneration of the corneal epithelium is driven by unipotent stem cells located at the limbus.1 However, when corneal epithelial regeneration is compromised due to LSCD, conjunctival and vascular cells migrate onto the corneal surface, which is often accompanied by destruction of the basement membrane, non-healing corneal epithelial defects, inflammation and scarring.1 The surgical therapy includes repeated abrasions, amniotic membrane transplantation and limbal stem cell transplantation depending on the extent of LSCD.2–5 For unilateral involvement, multiple transplantation techniques utilising autologous tissue have been developed, including conjunctival limbal autografts, cultivated limbal epithelial transplantation (CLET), and simple limbal epithelial transplantation (SLET). With all these techniques, excellent results have been achieved.6–9 Among these methods, SLET is cost-effective, surgically and logistically easy to perform and does not require much limbal tissue.8 However, in bilateral cases of LSCD, all transplantation methods can only be applied by using allogenic limbal tissue. The successful use of allogenic tissue for SLET was first reported in a patient with bilateral LSCD after chemical injury in 2013.10 Recently, Iyer et al also described SLET utilising allogenic simple limbal epithelial transplantation (alloSLET) as a useful technique to achieve rapid epithelialisation in severe chemical injuries.11 The present study is the first to evaluate the outcome of alloSLET utilising cadaveric donor tissue from cultured corneoscleral buttons in non-healing corneal epithelial defects with LSCD in patients with failed epithelialisation of the corneal surface.
Methods
This retrospective, interventional case series was performed at the Department of Ophthalmology, University Medical Center of the Johannes Gutenberg University Mainz. According to local law (‘Landeskrankenhausgesetz’ §36, §37), no ethical approval was required for this retrospective analysis.
Patients
Medical records of patients who had been treated for persistent corneal epithelial defects with alloSLET between September 2017 and October 2018 were reviewed. The inclusion criteria were (1) patients, who had undergone alloSLET for a non-healing epithelial defect with a history of failed epithelialisation due to total LSCD mostly after at least one superficial keratectomy with amniotic membrane transplantation. Of note, in all patients no autologous tissue was available either because of bilateral involvement (n=10), because the patient was afraid to worsen the situation in the other eye and therefore refused biopsy (n=3) or because the patient had only one eye (n=1). (2) A postoperative follow-up time of at least 12 months. All patients had scheduled examinations at the Department of Ophthalmology of the University Medical Center of the Johannes Gutenberg University Mainz 1, 3, 6 and 12 months after surgery, which included assessment of best corrected visual acuity (BCVA), slit lamp examination and fluorescein staining. Patients were also regularly visited by their local ophthalmologists and referred to our clinic in case of any questions or abnormalities.
The primary outcome measure was complete epithelialisation of the corneal surface. Secondary outcome measures were BCVA and postoperative side effects due to surgery or medical therapy. Patient’s characteristics are presented in table 1 and donor’s characteristics are presented in table 2.
Patient characteristics
Donor characteristics
Surgery
All surgeries were performed by the same experienced surgeon (AG), 12 under general anaesthesia and 2 under local anaesthesia. First, a 360° peritomy was conducted and the fibrovascular pannus removed by using a hockey knife and eye scissors. If necessary, bleeding was controlled by gentle cauterisation. In patients with deep corneal ulcers, penetrating keratoplasty (PK) was performed after removal of the pannus. After the penetrating corneal graft had been sewn in, its epithelium was removed by a hockey knife and a tying forceps. Next, a cryopreserved amniotic membrane graft (Eye Bank of Rhineland-Palatinate, Mainz, Germany) was placed stroma side down on the corneal stroma and bare sclera, secured with fibrin glue
(TISSEEL, Baxter Deutschland GmbH, Unterschleißheim, Germany) and was additionally secured with eight interrupted sutures (10.0 vicryl, Johnson & Johnson, Norderstedt, Germany) to the episclera. Subsequently, a 2 mm wide strip of donor limbal tissue was dissected 360° from the corneoscleral button by using a trephine, scalpel, Colibri forceps and eye scissors. Corneal stroma and scleral tissue were removed as much as possible by using spring and Vannas scissors. Next, the limbal tissue was cut into 50 to 80 pieces with Vannas scissors and placed on the periphery and mid-periphery of the amniotic membrane sparing the visual axis. The limbal tissue pieces were then secured with fibrin glue (TISSEEL, Baxter, Unterschleissheim, Germany). Three minutes later, a bandage contact lens of 20.5 mm diameter (Megasoft, Oculentis, Berlin, Germany) was placed on the ocular surface and left for 1 month. After 1 month, the contact lens was replaced by a new one for 2 more months. After 3 months, the contact lens was removed.
Postoperative treatment
Postoperatively, preservative-free 0.5% levofloxacin eye drops (Oftaquix sine, Santen GmbH, Munich, Germany) were applied four times daily until the contact lens was removed. Preservative-free 0.13% dexamethasone eye drops (Dexa EDO, Bausch & Lomb, Berlin, Germany) were administered six times daily for 4 weeks and then monthly reduced by one drop. After 5 months, patients remained on a dropping frequency of 1 to 2 times daily. Moreover, preservative-free lubricating eye drops containing 3% trehalose and 0.15% hyaluronic acid (Thealoz Duo, Thea Pharma GmbH, Berlin, Germany) were administered at least six times daily. Systemic therapy included mycophenolate mofetil (Cell-Cept, Roche Pharma AG, Grenzach, Deutschland) 1 to 2 g daily peroral and prednisolone (Dekortin H, Merck KGaA, Darmstadt, Germany) peroral at an initial dose of 1 mg/kg body weight, which was reduced within 1 month to 5 mg daily in total (table 1).
Statistical methods
Statistical analysis was performed using GraphPad Prism V.6 (GraphPad Software Inc, La Jolla, California, USA). Survival probability of alloSLET grafts was computed by Kaplan-Meier survival analysis. For comparisons of Kaplan-Meier curves, the Log-rank (Mantel-Cox) test and the Gehan-Breslow-Wilcoxon test were used. Postoperative BCVA was compared with preoperative values by a Wilcoxon matched-pairs signed rank test. Because BCVA was compared at each of the four postoperative time points with the preoperative time point, which resulted in four comparisons, the significance level α was adjusted by the Bonferroni correction to 0.0125. Comparisons of graft characteristics between patients who developed epithelial defects postoperatively and patients who maintained a stable corneal epithelium were conducted by the Mann-Whitney test. For these comparisons, the significance level was set 0.05.
Results
Fourteen eyes of 14 patients (9 males and 5 females), who received alloSLET due to persistent corneal epithelial defects because of total LSCD were identified. Thirteen of 14 cases had a history of one or more superficial keratectomies with amniotic membrane transplantations. The mean patients’ age was 59.4±13.6 years (range 43–88). The surgery was uneventful in all cases. In figure 1, the preoperative situation and postoperative pictures are presented for some patients. Thirteen (92.9%) of 14 patients had an epithelialised corneal surface 1, 3 and 6 months after surgery. In one patient with graft-versus-host disease and bilateral LSCD, who frequently lost his bandage contact lens, the corneal surface was not epithelialised 1 month after surgery. Due to progressive corneal melting, PK had to be performed and the patient was scheduled for tibia-keratoprosthesis. Twelve months after alloSLET, 10 (71.4%) of 14 patients had a completely epithelialised corneal surface (figure 2A). The underlying reasons for LSCD in the three patients, in whom the corneal grafts failed 7, 8 and 11 months after surgery, respectively, were chemical burn, blast injury and radiotherapy due to lymphoid hyperplasia of the orbit. The reason was chronic graft failure rather than acute immune rejection, because no circumcorneal congestion and no tortuosity of perilimbal blood vessels was seen in these patients. Notably, there was no different graft failure rate between patients, who received alloSLET only and patients who received alloSLET with PK. In both subgroups, 5 (71.4%) of 7 patients had an epithelialised corneal surface 12 months after surgery (figure 2B). In the whole patient collective, BCVA was markedly better 6 and 12 months after surgery compared with preoperative values (figure 3A). Twelve months postoperatively, 8 (57.1%) of 14 patients displayed an improved BCVA. Of note, no significant improvement of BCVA was observed when looking separately at the subgroups of patients who received alloSLET only (figure 3B) and alloSLET with PK (figure 3C). This is likely due to the low patient numbers (n=7 in each subgroup) and the great variability of visual outcome.
Slit-lamp photographs of selected cases showing the preoperative situation (a), the result 1 day (b), 3 months (c), 6 months (d) and 12 months (e) after alloSLET. 1a–e: patient with LSCD due to chemical eye injury, who received alloSLET only. When the cornea cleared up, a mature cataract became visible, which was possibly caused by prolonged steroid use in the past and was removed by cataract surgery 4 months after alloSLET. 2a–e: patient with LSCD due to chemical eye injury, who received alloSLET with PK. Of note, in this patient the limbal tissue bits moved to the superior temporal periphery after surgery. However, the corneal surface epithelialised and remained so even after 12 months. 3a–e: patient with LSCD following radiotherapy, because of lymphoid hyperplasia of the orbit. Twelve months following alloSLET with PK (3e), a central corneal epithelial defect with stromal opacification was visible on the fluorescein-stained graft. alloSLET, allogenic simple limbal epithelial transplantation; LSCD, limbal stem cell deficiency; PK, penetrating keratoplasty.
Kaplan-Meier analysis of stable corneal surface survival. (A) Survival analysis of all treated eyes (n=14). One year after surgery, 71.4% of eyes had an epithelialised corneal surface. (B) Eyes subjected to alloSLET only (n=7) had a similar survival curve as eyes treated by alloSLET with PK (n=7). Twelve months after surgery, the corneal surface was epithelialised in 71.4% of eyes from both subgroups. alloSLET, allogenic simple limbal epithelial transplantation; PK, penetrating keratoplasty
BCVA preoperatively and 1, 3, 6 and 12 months after alloSLET. in the whole patient collective (n=14), marked improvement of BCVA was observed 6 and 12 months after surgery (A). However, in the subgroups of patients, who received either alloSLET only (n=7, B) or alloSLET with PK (n=7, C) no significant improvement was seen. Data are presented as box-and-whisker plots including all individual points (**p<0.01). alloSLET, allogenic simple limbal epithelial transplantation; BCVA, best corrected visual acuity; PK, penetrating keratoplasty
In the present study, only two corneal grafts that served for alloSLET have been HLA-matched (Human leukozyte antigen-matched). One of them failed 7 months after transplantation. To investigate whether graft characteristics or culturing time predisposed to graft failure, we analysed the donor’s age, death-to-harvest time, endothelial cell density, total cultivation time and cultivation time in medium 1 and medium 2 for grafts which survived and for those that failed, respectively. Notably, none of these factors appeared to predispose to graft failure (figure 4). In 3 (21.4%) of 14 patients, postoperatively dislocated pieces of limbal tissue have been found in the centre of the cornea, possibly affecting vision. However, the central pieces resolved after 3 to 6 months.
Factors of donor tissue that may have reduced the quality of the transplanted graft pieces. Donor age (A), time between the donors’ death and harvesting of the graft tissue (B), endothelial cell density (C), total time of graft tissue cultivation (D), cultivation time in medium 1 (E) and cultivation time in medium 2 (F) are presented for the grafts that survived (n=10) and for those that failed (n=4). Of note, graft tissue from patients who experienced epithelial defects after alloSLET had similar characteristics as graft tissue from patients with intact epithelium. alloSLET, allogenic simple limbal epithelial transplantation.
Discussion
The present study demonstrates the successful treatment of non-healing corneal epithelial defects caused by total LSCD of different origin in cases where no autologous tissue was available by alloSLET from cadaveric cultured donor corneoscleral buttons.
All (92.9%) but one patient had an epithelialised corneal surface 6 months after surgery and 10 (71.4%) of 14 patients displayed an epithelialised corneal surface 12 months after surgery.
Unilateral LSCD can be treated with high success rates by different transplantation techniques utilising small amounts of autologous limbal tissue from the fellow eye.8 9 11–13 However, in bilateral LSCD these techniques can only be applied by using allogenic limbal tissue. Previous studies utilised allogenic tissue for keratolimbal allografts (KLAL) or CLET with variable outcomes. For example, the success rate of limbal allografts, defined as intact corneal epithelium, was reported to vary between 33% and 83% in non-HLA-matched grafts and between 75% and 100% in HLA-matched grafts.14–21 In the present study, only two corneal grafts that served for alloSLET have been HLA-matched. Although it may be favourable to use HLA-typed grafts for alloSLET to reduce the risk of immune rejection, the waiting time for HLA-matched tissue may take even years. Hence, utilisation of non-typed graft material may be preferable with regard to visual rehabilitation compared with the alternatives, such as conservative treatment or covering the cornea with a conjunctival flap or oral mucosa.22–24
Regarding the use of allogenic limbal tissue for CLET, Shimazaki et al performed a study with a mean follow-up time of 127 weeks. The authors reported on a stable corneal epithelium in 86% of recipients of autologous tissue, in 63% of recipients of tissue from living relatives and in 42% of recipients who received tissue from cadaveric donor eyes.25 Also, Pauklin et al treated patients with total or partial LSCD of different origin by CLET using either autologous or allogenic tissue expanded on intact amniotic membrane. After a mean follow-up time of 28.5 months, an epithelialised corneal surface was observed in 77% of patients after autologous CLET and in 50% of patients after allogenic CLET.26 In another study that utilised allogenic limbal tissue for CLET in patients with ocular burns, complete success defined as the absence of recurrent symblepharon formation was achieved in 50%, partial success in 31.3% and failure in 18.8% of the patients.27
Because limbal cells obtained from cadaveric donor eyes were reported to have a reduced proliferation rate in vitro and a blunted epithelialisation rate in vivo, harvesting tissue from living donors might be preferable.25 28
On the other hand, cadaveric donor limbal tissue is much easily available and its use allows for transplanting more tissue, which may compensate for the poorer cell viability.
SLET has the advantage of being a cost-effective, single-staged procedure which as opposed to CLET does not require a highly equipped laboratory to obtain CLET.8 Iyer et al have recently reported on the use of cadaveric donor tissue for alloSLET to accelerate corneal wound healing in the acute stage of severe chemical injuries. The mean time to complete epithelialisation was 22.5±9.14 days, which was achieved in 17 (94.11%) of 18 eyes.11 Moreover, very recently, Shanbhag et al reported on their experience in 16 eyes that underwent live-related alloSLET and 14 eyes that were subjected to cadaveric alloSLET.29 Successful outcomes were reported in 14 (87.5%) of 16 eyes in the live-related donor group and in 11 (78.6%) of 14 eyes in the cadaveric donor group 1 year postoperatively and at the final follow-up, which was 28 months (range 13–66 months).29 Regarding the use of cadaveric donor tissue for alloSLET, Shanbhag et al suggested to use fresh tissue, preferably before 48 hours from the time of harvesting, and tissue from donors aged not more than 60 years.29
Although utilisation of fresh tissue may be preferable, a time interval of less than 48 hours is barely feasible in many eye banks due to logistical delays, such as the waiting time for microbiological or serological results. In the present study, we used limbal tissue from cultured corneoscleral buttons with a cultivation time ranging from 11 to 29 days. Also, with an age range from 60 to 95 years our corneal donors were much older than recommended by Shanbhag et al. Remarkably, our postoperative 1 year success rate was quite close to that reported by Shanbhag et al (10 of 14 eyes vs 11 of 14 eyes).29 Notably, neither the age of the corneal donor, the death-to-harvest time nor the time in culture appeared to have a major impact on a successful or unsuccessful epithelialisation of the corneal surface. Hence, our study suggests that cultured tissue from patients older than 60 years is also suitable for alloSLET.
It has to be pointed out that alloSLET should not be regarded as a method that may compete with transplantation techniques utilising autologous limbal tissue, including traditional SLET, but rather as a therapy reserved for patients, who do not qualify for autologous limbal tissue transplantation. The advantage of alloSLET over allogenic CLET is being a single-stage procedure, its cost-effectiveness and the availability in clinics that have no laboratory to expand limbal tissue ex vivo. The presumed advantages of alloSLET over KLAL are the simplicity of the surgical procedure, which is reflected in a high success rate of traditional SLET even when performed by less experienced surgeons,12 no need of lamellar excision of limbal or scleral tissue to fit the graft and, thus, potential reversibility and easy repeatability of the procedure. Moreover, by placing the limbal tissue on the corneal surface, direct contact with blood vessels is avoided, which may reduce the risk of immunological rejection. Of note, in our study we transplanted much more limbal tissue than typically used for autologous SLET, because we wanted to compensate for the reduced proliferation rate of limbal cells from cadaveric donor eyes. However, some of the limbal allograft bits dislocated to the periphery and straddled the limbus, which may have increased the risk of immune reaction and, thus, reduced the potential benefit of alloSLET. Moreover, in three eyes the pieces of glued limbal tissue moved into the corneal centre, which probably prolonged visual recovery. Hence, the optimal amount of tissue to be transplanted enabling a high probability of long-term corneal surface recovery and reducing the side effects of tissue dislocation remains to be determined.
A disadvantage of alloSLET is the need of a bandage contact lens during the first postoperative weeks or months, which is at increased risk to be lost when the patient rubs the eye, has nystagmus or an abnormal lid or corneal anatomy. In such cases the use of KLAL would probably be a better option.
In summary, alloSLET utilising cadaveric limbal tissue from cultured donor corneoscleral buttons is an effective treatment option for corneal epithelial defects associated with LSCD when autologous tissue is not available. The primary outcome of the study was an epithelialised corneal surface, which was seen in 71.4% of eyes 12 months after surgery. Moreover, visual acuity markedly improved following alloSLET and side effects were minor.
References
Footnotes
Contributors JCR, NP and AG conceived the study. JCR, AM and JWP collected and collated the data. JCR and AG performed data analysis. All authors have contributed equally in the manuscript preparation.
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 Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement All relevant data to this study are included in the article. If further information are required please contact us (jana.riedl@unimedizin-mainz.de).
Linked Articles
- At a glance