Article Text
Abstract
Purpose To conduct a systematic review on outcomes of three different techniques of autologous limbal stem cell transplantation (LSCT): conjunctival-limbal autografting (CLAu), cultivated limbal epithelial transplantation (CLET) and simple limbal epithelial transplantation (SLET), in unilateral limbal stem cell deficiency (LSCD).
Methods Literature searches were conducted in MEDLINE (Ovid), Embase, Web of Science and Cochrane Central Register. Standard systematic review methodology was followed using Meta-analysis of Observational Studies in Epidemiology guidelines. Studies with a sample size of more than 10 eyes were included. The primary outcome measure of efficacy was restoration of a completely epithelised, stable and avascular corneal surface (anatomical success). The secondary outcome measure of efficacy was improvement in best-corrected visual acuity of two-lines or greater (functional success).
Results The review identified 22 non-comparative case series, which included 1023 eyes. Ocular burns were the major (88%) indication for surgery. Overall, at a median postoperative follow-up of 1.75 years, autologous LSCT for unilateral LSCD showed anatomical and functional success rates of 69% and 60%, respectively, without any serious adverse events in the donor eye. The follow-up duration and indications for surgery were comparable across all groups (p>0.05). The anatomical and functional success rates of SLET (78%; 68.6%) and CLAu (81%; 74.4%) were comparable, and significantly better than those of CLET (61.4%; 53%; p=0.0048).
Conclusion Autologous LSCT is a safe and effective treatment for unilateral LSCD. In the absence of randomised controlled trials, existing evidence clearly suggests that clinical outcomes are better with SLET and CLAu as compared with CLET.
- limbal stem cell transplantation
- conjunctival limbal autograft
- cultivated limbal epithelial transplantation
- simple limbal epithelial transplantation
- unilateral limbal stem cell deficiency
- ocular chemical burns
Statistics from Altmetric.com
- limbal stem cell transplantation
- conjunctival limbal autograft
- cultivated limbal epithelial transplantation
- simple limbal epithelial transplantation
- unilateral limbal stem cell deficiency
- ocular chemical burns
Introduction
The limbus contains stem cells which continually replenish the corneal epithelium in the physiological state.1 Limbal stem cell deficiency (LSCD) arises in eyes where these epithelial stem cells are damaged.2 3 In LSCD, the corneal epithelium is replaced by the conjunctiva, leading to loss of corneal clarity and impairment of vision. Corneal transplantation alone is ineffective in LSCD and invariably results in repeated surface breakdowns and epithelial healing issues, ultimately leading to graft failure.4 Fortunately, limbal stem cell transplantation (LSCT) can restore the normal epithelial phenotype and re-establish a stable corneal surface.5 In patients with unilateral LSCD, the healthy fellow eye can serve as the source for donor limbal tissue, which makes the procedure autologous and obviates the need for any systemic immunosuppression.
While the basic concept of autologous LSCT has remained unchanged over the last three decades, not only have three different surgical approaches evolved with time, but they also continue to be practised concurrently. The original technique of LSCT, conjunctival-limbal autografting (CLAu), was described by Kenyon and Tseng in 1989 and involves direct transplantation of two 3-clock-hour long conjunctival-limbal lenticules from the healthy eye onto the diseased limbal bed.5 Subsequently in 1997, Pellegrini et al described cultivated limbal epithelial transplantation (CLET), where a 2 mm×2 mm limbal biopsy from the donor eye is expanded ex-vivo in the laboratory for 10–14 days into a sheet of epithelial cells and then transplanted onto the surface of the affected eye.6 Finally, Sangwan et al in 2012, described the technique of simple limbal epithelial transplantation (SLET) where a strip of donor limbal tissue from the superior limbus, similar to CLET, is obtained from the healthy eye, but instead of ex-vivo expansion it is divided into small pieces and transplanted directly onto the affected cornea for in-vivo expansion over a human amniotic membrane graft.7 Although the long-term clinical outcomes of all three LSCT-techniques have been reported, there is no clear consensus on whether one technique is superior in any way to the others. Currently this choice is based on the surgeon’s individual preference and familiarity, rather than evidence. To address this issue, the authors conducted a systematic review on the clinical outcomes of the three techniques of autologous LSCT, namely CLAu, CLET and SLET, in eyes with unilateral LSCD.
Methods
Search strategy
This systematic review was conducted as per the Meta-analysis of Observational Studies in Epidemiology guidelines.8 The systematic review protocol was registered at the Prospective Register for Systematic Reviews (PROSPERO; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=89913). The search strategy for this review is reported in online supplementary appendix A. The search was conducted during July 2018 in the following electronic databases: MEDLINE (Ovid), Embase, Web of Science, and the Cochrane Central Register. PROSPERO was searched for ongoing or recently completed systematic reviews. All reports published before 1 July 2018 were screened. No specific dates or years were used as time limits. No language restrictions were used. Reference lists were scanned to identify additional applicable studies.
Supplemental material
Inclusion criteria
Reports in English language were considered for analysis. Reports where outcomes of the following interventions: CLAu, CLET and SLET could be discerned in >10 patients with total unilateral LSCD were included. Reports where repeat LSCT was performed for eyes in which previously performed LSCT had failed were also included.
Exclusion criteria
Letters, conference abstracts, case reports, review articles, editorials and animal studies were excluded. Other excluded studies were: (1) those which focused on the outcomes of penetrating keratoplasty after LSCT without describing the outcomes of the initially performed LSCT; (2) those focusing on outcomes of LSCT in partial LSCD alone; (3) those that included both partial LSCD (where limbal biopsy was obtained from the same eye) and total LSCD, but ≤10 eyes had total LSCD; (4) reports that included both autologous and allogeneic LSCT, but where outcomes specific to the autografts could not be discerned; (5) reports where at least one of the two primary outcome measures could not be discerned; (6) those with a mean follow-up duration of <6 months.
Special considerations
Some studies classified outcomes as success, partial success and failures. However, the criteria used to define partial success were consistent with that of failure in other studies. For the sake of consistency, partial successes were therefore counted as failures. In studies which reported the outcomes of the same procedure being repeated, only the outcomes of the primary procedure were considered for analysis. Measures were taken to avoid including multiple studies reporting outcomes of the same cohort of patients at different time intervals.
Article selection and data extraction
Of all the studies that were suitable for inclusion, the titles and abstracts were screened by two authors (SSS and NN). Disagreements were resolved by consensus of all authors. Selected full-text articles were retrieved and assessed for eligibility. From the full-text studies that were suitable for inclusion, information was recorded on a data extraction form. Data recorded included study design, study population, aetiology of LSCD, interventions, clinical outcomes, mean follow-up period and complications. A descriptive synthesis was undertaken. The primary outcome measure of efficacy was the restoration of a completely epithelised, stable and avascular corneal surface (anatomical success). The secondary outcome measure of efficacy was a two-line or greater improvement in best-corrected visual acuity (BCVA) after intervention (functional success).
Quality assessment
To rate individual studies based on study design, a scale based on the Scottish Intercollegiate Guidelines Network (https://www.sign.ac.uk/assets/qrg50.pdf, last accessed on 31 December, 2018) was used. Two reviewers (SSS, PRD) independently assessed the quality of all the included studies using a validated checklist consisting of 14 questions (online supplemental table). Any disagreements were resolved by consensus or arbitration by the senior author (SB). This checklist was developed by using a combination of two quality appraisal checklists specifically devised for interventional case series.9 10 Items that were applicable to this study were used. A score of 1 was assigned when the answer to a quality question was clearly affirmative. The score for all studies were collated.
Supplemental material
Statistical analysis
MedCalc (V.11.4.3.0, MedCalc Software, Mariakerke, Belgium) statistical software was used for data analysis. The anatomical and functional success rates were reported as proportions (percentages) with 95% CIs. The study quality, sample size, indication for surgery and follow-up duration between the groups were compared using the one-way-analysis of variance and Student-Newman-Keuls test. The comparisons in success rates between the different techniques were performed using the Pearson’s χ2 test. A p value <0.05 was considered statistically significant.
Results
Selection of studies for review
A total of 5588 articles of potential interest were identified by the original literature search (figure 1). No similar ongoing or recently completed systematic reviews were found on PROSPERO. After removal of duplicated records, 3422 records were screened. After reviewing the titles and abstract, 3366 articles were excluded, and 56 full-text articles were assessed for eligibility. The eligibility criteria were met by a total of 22 reports and these were included in this systematic review.5 11–31 Full-text articles that did not fulfil the eligibility criteria included: (1) multiple studies on CLAu in cases of pterygium; (2) studies with a sample size of ≤10 eyes with total LSCD32–34; (3) a study where the outcomes of autografts and allografts could not be differentiated35; (4) one study which reported outcomes of autologous LSCT at a mean follow-up of <6 months.36 Several groups reported outcomes in the same cohort of patients at different time intervals, and hence, only the studies with longest duration of follow-up were included.37–41
Study designs and quality assessment
There were no randomised controlled trials; all studies were non-comparative interventional case series. Descriptive information for each study is presented in table 1. All 22 studies included in this systematic review were rated and the quality assessed (table 2 and figure 2). When a quality assessment question was graded as “yes”, it represented a positive measure of quality. Of the 22, five (23%) studies collected data prospectively, one (5%) study was multicentric, 13 (59%) studies clearly mentioned the recruitment period, 18 (82%) studies mentioned the characteristics of the patients. The inclusion and exclusion criteria were clearly mentioned in 14 (64%) studies. None of the 22 studies had participants at a similar stage of disease preoperatively. The intervention of interest (along with the number of clock hours of limbal biopsy) was clearly described in 17 (77%) studies. Three (14%) studies clearly mentioned that the surgeries were performed by an experienced surgeon. Additional postoperative surgical interventions were clearly described in 18 (82%). Relevant outcomes measures were clearly mentioned in 18 studies (82%). Only one (5%) study clearly mentioned that the outcome assessors were blinded regarding the intervention. All-important outcome measures were considered and evaluated in 15 (68%) studies. The outcomes were evaluated at a follow-up of 6 months or more in 16 (73%) studies (in six studies, there were some individual patients with follow-up of <6 months, but the overall mean follow-up was ≥6 months). Adverse events were reported in 19 (86%) studies. The 10 studies where CLAu was performed had an average quality score of 5.6/14 (40%; 95% CI 32% to 49%). The eight studies where CLET was performed had an average quality score of 7.6/14 (54.4%; 95% CI 45% to 64%) while the four studies where SLET was performed had an average quality score of 10/14 (71.4%; 95% CI 58% to 83%).
Efficacy of CLAu
The outcomes of CLAu were analysed in 189 eyes across 10 studies.5 11–19 Eight studies with 162 eyes provided data regarding aetiology, and 83% (135/162) eyes had LSCD secondary to chemical burns. The primary outcome measure of a stable corneal surface at the end of the follow-up period was mentioned in 168 eyes across nine studies. At a mean follow-up period of 1.56 years, 81% (136/168, 95% CI 74.2% to 86.6%) eyes had a stable, fully epithelised and avascular surface. The secondary outcome measure of a two-line improvement in BCVA could be discerned in 117 eyes across six studies. At a mean follow-up period of 1.65 years, 74.4% (87/117, 95% CI 65.5% to 82%) eyes had a two-line improvement in BCVA.
Efficacy of CLET
The outcomes of CLET were analysed in 581 eyes across eight studies.20–27 All eight studies provided data regarding aetiology, and 90% (521/581) eyes had LSCD secondary to chemical burns. The primary outcome measure of a stable corneal surface at the end of the follow-up period was mentioned in all eight studies. At a mean follow-up period of 2.9 years, 61.4% (357/581; 95% CI 57.4% to 65.4%) eyes had a stable, epithelised, avascular surface. The secondary outcome measure of a two-line improvement in BCVA was mentioned in 425 eyes across six studies. At a mean follow-up period of 2.4 years, 51.5% (219/425; 95% CI 46.7% to 56.4%) eyes had a two-line improvement in BCVA.
Efficacy of SLET
The outcomes of SLET were analysed in 253 eyes across four studies.28–31 All four studies provided data regarding aetiology, and 93% (235/253) eyes had LSCD secondary to chemical burns. The primary outcome measure of a stable surface at the end of the follow-up period was mentioned in all four studies. At a mean follow-up period of 1.48 years, 78% (197/253, 95% CI 72.2% to 82.8%) eyes had a stable, epithelised, avascular surface. The secondary outcome measure of a two-line improvement in BCVA was mentioned in 223 eyes across three studies. At a mean follow-up period of 1.2 years, 68.6% (153/223; 95% CI 62.1% to 74.6%) eyes had a two-line improvement in BCVA.
Comparison of primary and secondary measures (CLAu vs CLET vs SLET)
Overall, at a median postoperative follow-up of 1.75 years (range, 1–7.6 years), 68.9% (690/1002, 95% CI 65.9% to 71.7%) eyes maintained a stable, epithelised, avascular cornea, and 60.8% (459/755, 95% CI 57.2% to 64.3%) eyes gained at least two-lines of BCVA. The overall follow-up (p=0.358), and indications for surgery (p=0.089) were comparable across all three groups. However, study quality (p=0.003) and sample size (p=0.039) were significantly lesser in the CLAu group. Statistically, SLET and CLAu had significantly better outcomes compared with CLET, both in terms of anatomical (p=0.0048) and functional success (p≤0.0001). The anatomical success rates were almost identical between SLET and CLAu, while the functional success rates were marginally better with CLAu, although not statistically significant (p=0.27). The results are summarised in table 1 and figure 3.
Safety in the recipient eye
The reported complications of autologous LSCT are listed in table 1. In most studies, the most common early postoperative complication was haemorrhage under the amniotic membrane graft. Late postoperative complications were either recurrence of LSCD or persistent epithelial defects. The latter rarely led to secondary complications such as perforation or microbial keratitis.
Safety in the donor eye
In 6/10 studies on CLAu, donor-site complications were looked for but none were noted. In 3/8 studies on CLET, donor-site complications were looked for and again none were noted. In all four studies on SLET, the donor-eye was evaluated postoperatively and findings were noted, but no serious adverse events were reported.
Discussion
In the absence of randomised controlled trials, systematic reviews of non-comparative case series form the strongest pillars of evidence-based medicine. This systematic review compared the anatomical and functional outcomes of three different techniques of autologous LSCT for unilateral LSCD. After a thorough literature review, the authors identified 22 non-comparative case series that included more than 10 cases with at least 6 months of mean postoperative follow-up. Most of the cases covered in this systematic review had LSCD secondary to chemical burns. All cases underwent autologous grafts from the unaffected fellow eye without any long-term topical or systemic immunosuppressive medications. Similarities in indication for surgery and follow-up duration across the three groups provided for a uniform platform to compare the individual techniques. The findings of this systematic review clearly suggest that: (1) irrespective of the technique used, autologous LSCT is effective in treating unilateral LSCD and safe for the donor eye; (ii) SLET and CLAu have significantly better anatomical and functional success rates compared with CLET.
The strengths of this study include a robust methodology, with rigorous inclusion and exclusion criteria, adopted to select for reliable studies within each group with adequate numbers and follow-up. However, there are caveats. Since ocular burns in general and chemical burns in particular, were the major indications for surgery across all studies, the results of this review should not be extrapolated to other indications of autologous LSCT. With regards to both CLAu and SLET, the studies within each group used similar techniques and therefore minor variations in the surgical methods were unlikely to affect the validity of the comparisons. For CLET, however, varying cultivation protocols were used by different groups, which may have influenced the overall outcomes. Nevertheless, no individual study on CLET, irrespective of the laboratory protocols, reported success rates that were better than those noted for CLAu or SLET. Finally, the success rates observed with CLAu should be placed in the context of the fact that this group had the smallest sample size and lowest study quality.
The overarching goal of this systematic review was to help the treating surgeon decide which technique of autologous LSCT to use when confronted with a patient with unilateral LSCD. Although this review does not answer the question directly, it does provide direction. The results indicate that those practising CLAu or SLET can continue to do so and do not need to “upgrade” or shift to a different technique. However, the high laboratory costs for CLET may be difficult to justify considering the finding that it does not provide any proven clinical benefit over CLAu or SLET. Intuitively, SLET has the advantage over CLAu of needing less tissue from the donor eye. Although there is no data regarding efficacy based on the severity of injury, the authors recommend SLET for LSCD cases with milder conjunctival involvement (lesser grades of symblepharon) and CLAu for LSCD with more severe conjunctival involvement.42 Also, in circumstances where the cost or availability of the amniotic membrane and/or fibrin glue may be limiting factors, CLAu may be preferred over SLET. However, despite the poorer clinical results with ex-vivo cultivation, it will continue to have relevant research and clinical applications. For example, CLET can be combined with gene therapy for congenital causes of LSCD such as epidermolysis bullosa, which cannot be treated by SLET or CLAu.43 44 Therefore, each of the three procedures may be preferred in specific situations, rather than one procedure being the recommended standard of care.
In conclusion, this systematic review of 22 studies including 1023 eyes aimed to compare the clinical outcomes of three different techniques of autologous LSCT for the treatment of unilateral LSCD. The findings of the review clearly indicate that autologous LSCT, irrespective of the technique used is a safe and effective therapy for this condition; and that SLET and CLAu have better long-term anatomical and functional success rates in comparison to CLET.
Acknowledgments
The authors thank Paul Bain, Reference Librarian at Countway Medical Library, Harvard Medical School for providing assistance with the literature search.
References
Footnotes
Contributors The corresponding author states that authorship credit of this manuscript was based on (1) substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data; (2) drafting the article or revising it critically for important intellectual content; (3) final approval of the version to be published. All listed authors met conditions 1–3. All persons designated as authors qualify for authorship, and all those who qualify are listed. Each author has participated sufficiently in the work to take public responsibility for appropriate portions of the content.
Funding Hyderabad Eye Research Foundation, Hyderabad, India.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement All data relevant to the study are included in the article or uploaded as supplementary information.
Linked Articles
- At a glance