Article Text

Download PDFPDF
Management of limbal stem cell deficiency by amnion-assisted conjunctival epithelial redirection using vacuum-dried amniotic membrane and fibrin glue
  1. Harminder Singh Dua1,2,
  2. Darren Shu Jeng Ting1,2,
  3. Ahmed AlSaadi3,
  4. Dalia G Said1,2
  1. 1Academic Ophthalmology, School of Medicine, University of Nottingham, Nottingham, UK
  2. 2Department of Ophthalmology, Queen's Medical Centre, Nottingham, UK
  3. 3Department of Ophthalmology, Zayed Military Hospital, Abu Dhabi, UAE
  1. Correspondence to Professor Harminder Singh Dua, Ophthalmology, University of Nottingham, Nottingham, NG7 2UH, UK; harminder.dua{at}nottingham.ac.uk

Abstract

Purpose To study the outcome of a modified amnion-assisted conjunctival epithelial redirection (ACER) technique using vacuum-dried amnion (Omnigen) and fibrin glue for managing total limbal stem cell deficiency (LSCD).

Method A retrospective, interventional case series of all patients with total LSCD who underwent limbal stem cell transplant (LSCT) using the modified ACER procedure between 2016 and 2019. The outcome was defined as: (1) success: complete corneal re-epithelialisation without conjunctivalisation; (2) partial success: sub-total corneal re-epithelialisation with partial non-progressive conjunctivalisation sparing the visual axis and (3) failure: conjunctivalisation affecting the visual axis.

Results Ten patients (six men), with a mean age of 46.2±18.4 years, were included. The mean follow-up was 23.0±13.9 months. Causes of LSCD were chemical eye injury (30%), congenital aniridia-related keratopathy (30%), ocular surface malignancy (20%), Steven-Johnson syndrome (10%) and contact lens overuse (10%). 50% were bilateral. The time from diagnosis to ACER (for acquired causes) was 45.6±44.4 months. 80% of patients achieved a complete/partial success following ACER and 20% of patients required repeat LSCT. Auto-LSCT was associated with a significantly higher chance of success than allo-LSCT (p=0.048). The mean best-corrected-visual-acuity (logMAR) improved significantly from 1.76±0.64 preoperatively to 0.94±0.94 at final follow-up (p=0.009). Omnigen was available off-the-shelf stored at room temperature and its transparency enabled visualisation of the healing epithelium beneath.

Conclusion LSCT using the modified ACER serves as an effective ocular surface reconstruction technique in managing total LSCD and improving vision. Vacuum-dried amnion provides advantages of easy handling, transparency and storage at room temperature.

  • ocular surface
  • cornea
  • stem cells

Data availability statement

Data are available upon reasonable request. Individual patient data are stored in the hospital’s electronic system and can be retrieved on reasonable request.

Statistics from Altmetric.com

Request Permissions

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.

Introduction

Limbal stem cell deficiency (LSCD) is a challenging ocular surface condition that can result in significant ocular discomfort and visual impairment. It can be caused by a wide range of diseases, including chemical eye injury, long-term contact lens wear, ocular surface neoplasia, radiation keratopathy, drug toxicity, Steven-Johnson syndrome (SJS), ocular cicatricial pemphigoid and aniridia-related keratopathy, among others.1–6

The management of LSCD is dependent on the laterality, severity and extent of the disease, the underlying cause and the health status of ocular surface.7 Partial LSCD can be managed by sequential sectoral conjunctival epitheliectomy (SSCE) whereas total LSCD requires removal of fibrovascular pannus followed by ocular surface reconstruction with in vivo or ex vivo limbal epithelial cell expansion techniques.7–9 In in vivo expansion techniques, donor limbus (epithelial stem cells) tissue is obtained from the opposite unaffected eye (autografts) or from living relatives or cadaver donors (allografts) and transplanted directly to the affected eye. Depending on the donor tissue used, the procedures are termed as follows: conjunctival limbal autografts (CLAU) or conjunctival limbal allograft (CLAL), the latter could be living related or living non-related.7 The same tissue but including 0.5–0.75 mm of peripheral cornea would be termed CLPCoAU or CLPCoAL.10 CLAU, CLAL, CLPCoAU and CLPCoAL involve the transplantation of two pieces of 1–2 clock hours of limbus, usually superior and inferiorly, leaving 4–5 clock hours exposed on either side. During re-epithelialisation, while transplanted limbus-derived epithelial cells migrate circumferentially and centripetally to cover the corneal surface, conjunctival epithelial cells also migrate centripetally and cross the limbus to result in an admixture of conjunctival and corneal epithelial phenotype of cells on the corneal surface.11 12

Simple limbal epithelial transplant (SLET) is another in vivo expansion technique where small pieces of donor limbus (auto or allografts) are glued on the prepared corneal surface of the recipient eye.13 14 Ex vivo cultivation techniques employ epithelial cell sheets generated in vitro on a suitable substrate like amnion or fibrin (Holoclar), from small donor limbal biopsies. These pre-prepared cell sheets, rich in stem cells (holoclones) are transplanted (cultivated limbal epithelial transplant), from either auto or allo limbus tissue, to the affected cornea. In vivo expansion techniques are the most popular because of their comparatively low cost. In the UK the governing body, National Institute of Health and Care Excellence, requires that an in vivo expansion technique must be tried and Holoclar (the only licensed stem cell product) be used only if the former fails.

If, after removal of the fibrovascular pannus, the underlying corneal stromal surface is irregular and scarred, it can be covered by a circular disc of amniotic membrane, which provides a suitable substrate for the explant-derived epithelial cells to grow on, and becomes incorporated in to the corneal stromal tissue (graft or inlay).9 15 16 This is described as an in vivo expansion technique.9 After re-epithelialisation is complete, stable and stratified, improvement in vision is expected if the cornea is transparent. When the affected cornea is scarred, a corneal transplant procedure, either penetrating or deep anterior lamellar keratoplasty, is required as a second step towards visual rehabilitation.8 17

The donor limbal explants are usually between 1 and 2 clock hours wide and placed opposite each other on limbus of the recipient eye, usually at the 12 and 6 o’clock positions. This leaves large gaps nasally and temporally across which the recipient eye’s conjunctival cells can grow on the cornea and mix with the explant-derived corneal epithelial cells giving a less than satisfactory outcome. The conjunctiva-derived epithelial cells can be kept at bay by the method of SSCE.10 11 18 However, this is disadvantaged by the need for regular postoperative examination (eg, daily or alternate day), the potential risk of bleeding and ocular surface discomfort and/or pain.

To overcome the drawbacks of SSCE, we devised a technique called amnion-assisted conjunctival epithelial redirection (ACER),10 wherein the transplanted explants of limbal tissue are covered by a large patch (onlay) of cryopreserved amnion such that the nasal and temporal 4–5 clock hours of the edge of the membrane are tucked under and sutured to the peritomised and recessed conjunctiva, ensuring that any migration of the conjunctival epithelial cells occurs on the amniotic membrane while healing of the cornea beneath the amnion occurs from the explants, completely preventing admixture with conjunctival epithelium.

Since 2016, we further modified and refined our ACER technique using vacuum-dried amnion (Omnigen) and fibrin glue to reduce operating time and improve outcomes. Here, we report the method and outcomes of this modified technique for managing LSCD and discuss the advantages.

Materials and methods

This was a retrospective, interventional case series of all patients with total LSCD who underwent ocular surface reconstruction with limbal stem cell transplant (LSCT) using a modified ACER technique at the Queen’s Medical Centre, Nottingham, UK, between January 2016 and December 2019. All relevant data, including the history of the current complaints, the duration and nature of the disease or condition, preoperative and postoperative best-corrected-visual-acuity (BCVA), time taken for complete corneal healing postoperatively, any additional medical and surgical treatment, were collected from the medical case notes and were analysed. The work was approved as a clinical effectiveness study by the Clinical Audit and Effectiveness Department of Nottingham University Hospitals National Health Service Trust, Nottingham, UK (project number: 20-651C).

All patients underwent a complete clinical examination with the slit lamp microscopy, anterior segment optical coherence tomography (OCT) and in vivo confocal microscopy (IVCM). Patients with dry eye (Schirmer’s test (without anaesthesia) <5 mm) and inadequate lid closure were excluded, as previously described.8 9 The diagnosis of total LSCD was made based on clinical signs, including 360° involvement of the limbus and conjunctivalisation of the peripheral cornea, with or without involvement of the central cornea, presence of fibrovascular pannus with late staining covering corneal epithelium, columnar keratopathy and IVCM findings.1 19 20 In some cases, a variable area of a central island of healthy corneal epithelium was present despite the presence of total LSCD.21 Where a fundal view was not possible, ultrasound biomicroscopy was performed to assess the vitreous and anatomical status of the retina and choroid. Electrophysiological testing, including flash electroretinogram and flash visual evoked potential, was carried out to assess the optic nerve and retinal functions where visual acuity was ‘hand movements’ or less. Histological diagnosis was confirmed in all patients, either preoperatively by incisional biopsy (in cases of suspected ocular surface neoplasia) or postoperatively by the tissue removed during the surgery (ie, excisional biopsy). Only patients with potential for visual improvement were offered surgery. The patient and all potential allograft donor relatives were human leukocyte antigen (HLA)-typed and the best match was taken. The relationship of the donor to the patient is indicated in table 1.

Table 1

Summary of patients treated with limbal stem cell transplant (LSCT) using a modified amnion-assisted conjunctival epithelium redirection (ACER) technique for total limbal stem cell deficiency (LSCD)

Surgical technique

This modified technique was an adaptation from our previously described ACER technique.10 Under regional (sub-Tenon’s or peribulbar) or general anaesthesia, a 360° peritomy of the conjunctiva was performed. Any adhesions encountered were dissected or excised and the conjunctiva was recessed from the limbus. The edge of the fibrovascular pannus covering the cornea and limbus was dissected with a crescent blade (MicroSurgical Technology, Redmond, Washington, USA) from the periphery towards the centre until a distinct plane of cleavage was found. Working along this plane with a blunt instrument and incising any adhesions, the pannus was peeled off the stromal bed. Bleeding points were gently cauterised with a bipolar cautery.

LSCT was performed in the form of either CLPCoAU or CLPCoAL in all cases.10 Briefly, two limbal explants, comprising 2 clock hours of limbus, 3 mm of adjacent conjunctiva peripherally and 0.5–0.75 mm of peripheral cornea centrally, were excised from the 12 and 6 o’clock positions of the donor eye. The depth of conjunctival tissue was up to but not including the Tenon’s capsule and of the limbus and peripheral corneal stroma was between 100 and 150 µm. The explants were sutured to the respective positions of the recipient eye with two interrupted 10–0 monofilament nylon sutures placed diagonally, one at each end of the explant centrally and a long mattress suture, tangential to the limbus at the juncture of the conjunctiva with the limbus of the explant.

Omnigen500 (Vacuum-dried amniotic membrane in 25 mm diameter), epithelial side up, was placed on the recipient ocular surface, centred on the cornea. Two radial incisions were made in the periphery of the membrane on either side of the part covering the explants at 12 and 6 o’clock positions. The periphery of the membrane was attached to the sclera between nasal and temporal edges of the superior and inferior explants with fibrin glue (Tisseel; Baxter, Berkshire, UK). The corresponding edges of the recessed conjunctiva were then attached to the anterior surface of the amniotic membrane with fibrin glue such that it covered the peripheral 3–5 mm of the membrane and remained recessed from the limbus. The donor conjunctiva was attached to the edge of the recipient conjunctiva, superiorly and inferiorly, with fibrin glue. The small flap of amniotic membrane between the radial incisions was trimmed such that the limbal and peripheral corneal part of the explant remained covered by the membrane. In two cases where the corneal stromal bed was irregular, Omnigen80 (in 10 mm diameter disc of Omnigen) was affixed on the central cornea as an inlay graft prior to the transplantation of the larger overlay patch.

Healing of the corneal surface from limbal explant-derived epithelium was observed through the transparent Omnigen. When healing was complete, if the amnion had not spontaneously started to fall off, it was removed at the slit lamp under topical anaesthesia, by incising it at the junction with the conjunctival edge with a pair of disposable Vannas scissors and gently peeling it off the surface.

Topical preservative-free chloramphenicol drops (antibiotic) and dexamethasone drops (steroid) were given four times a day postoperatively. For allogeneic transplants, systemic immunosuppression with tacrolimus (Prograf; Astellas Pharma, Surrey, UK) 1 mg two times a day was initiated. Postoperative follow-up was at day 1, week 1, month 1, months 2, 4 and 6 and as required thereafter. Our regime was to continue systemic immunosuppression for 18 months after the operation or 18 months from the last rejection episode if the graft remains viable. This was empirically based on our experience wherein most rejection episodes for all high-risk corneal grafts and limbal grafts occur in the first 18 months.22 23

Outcome measures

The primary outcome of LSCT was divided into three categories: (1) success: complete corneal re-epithelialisation with explant-derived corneal phenotype epithelium, without any conjunctivalisation. (2) partial success: complete ocular surface re-epithelialisation with evidence of non-progressive conjunctivalisation of the peripheral cornea sparing the visual axis and (3) failure: conjunctivalisation of the cornea involving the visual axis. Improvement in vision was a secondary outcome of LSCT as the visual prognosis of some cases such as congenital aniridia was guarded due to optic nerve and retinal pathology.

Statistical analysis

Statistical analysis was performed using SPSS Statistics V.26 (IBM SPSS Statistics for Windows, Armonk, New York, USA). For descriptive and analytic purposes, the BCVA was converted from Snellen vision to logMAR vision. Counting fingers and hand motion were quantified as 1.9 logMAR and 2.3 logMAR, respectively.24 Comparison between groups was conducted using Pearson’s χ2 or Fisher’s Exact test where appropriate for categorical variables, and Student’s t-test or Mann-Whiney U test for continuous variables. Normality of data distribution was assumed if the skewness and kurtosis z-values were between −1.96 and +1.96 and the Shapiro-Wilk test p value was >0.05. All continuous data were presented as mean±SD. P value of ≤0.05 was considered statistically significant.

Results

Ten patients with total LSCD were included in this study; the mean age was 46.2±18.4 years (range 11–74 years), with a slight (60%) male preponderance (table 1). The mean follow-up duration was 23.0±13.9 months.

Nature and causes of LSCD

Five (50%) patients had total unilateral LSCD and another five (50%) patients had total bilateral LSCD. The causes of LSCD were chemical eye injury (3, 30%), congenital aniridia-related keratopathy (3, 30%), ocular surface malignancy (2, 20%), SJS (1, 10%) and contact lens overuse (1, 10%). According to the staging proposed in the global consensus report,1 all patients had a stage III LSCD (100% limbal involvement with associated involvement of the central 5 mm of the cornea). In both the malignancy cases, total LSCD developed as a result of complete invasion of conjunctival intraepithelial neoplasia (patient 7) and damage to the limbus due to previous radiotherapy and excision of conjunctival melanoma (patient 9). In the SJS case, LSCD was diagnosed during the chronic phase where there was presence of dry eye disease and lid margin keratinisation. The majority (9, 90%) of patients had a BCVA of 6/60 or worse. Two (20%) patients underwent fine needle diathermy and subconjunctival bevacizumab (Avastin) injection for treating corneal vascularisation 1–2 months before the ACER procedure.

Clinical outcomes

The mean time interval between the diagnosis of LSCD and ACER procedure (for acquired causes only) was 45.6±44.4 months (ranged, 18.6–144.7 months). Five (50%) patients underwent auto-limbal grafts, and five (50%) had living-related allo-LSCT. The duration for which Omnigen stayed on the ocular surface was 3.27±1.38 weeks. The mean time from LSCT to complete epithelialisation of the cornea was 1.14±0.97 months (ranged, 0.20–3.42 months). The outcome of ACER was considered a success in six (60%) patients (figures 1–3), a partial success in two (20%) patients and a failure in two (20%) patients. Auto-LSCT was associated with a significantly higher chance of success compared with allo-LSCT (p=0.048). All allo-LSCT cases were performed using grafts from a first-degree relative. At final follow-up, seven (70%) of all 10 patients achieved an improvement in the BCVA. The mean BCVA for all patients, including both congenital and acquired causes, significantly improved from 1.76±0.64 logMAR preoperatively to 0.94±0.94 logMAR at the final follow-up (p=0.009). However, for all successful and partially successful cases excluding the aniridia patients (n=6), the mean BCVA improved from 1.53±0.74 preoperative to 0.23±0.25 at final follow-up (p=0.004). In the eight cases where complete or partial success was achieved, the mean duration of complete surface epithelialisation was 4.96±4.20 weeks. Two of the three patients with aniridia were also on anti-glaucoma treatment, with one achieving partial success (patient 10) and the other experiencing treatment failure (patient 8). One patient with SJS (patient 3) and one with aniridia (patient 8) had an acute rejection of the limbal graft and corneal graft at 14 months and 3 months, respectively, postsurgery. This was followed by re-vascularisation and encroachment of fibrous growth on the cornea. Patient 3 suffered from corneal infection and perforation. Both these grafts failed despite additional oral steroid medication (SJS patient).

Figure 1

Diffuse unilateral papilloma/ocular surface squamous papilloma treated with autolimbal transplant by the ACER technique. (A) Clinical image at presentation. The papilloma with fronds of vessels is visible along the limbus and adjacent conjunctiva. (B) Preoperative image of the same eye after several years of treatment with mitomycin C and interferon alpha. (C) Day 1 postoperative. The transplanted autolimbal explants are visible through the Omnigen, at the 12 and 6 o’clock positions. Fine wrinkles are visible in the amniotic membrane. (D) Three days later, a sheet of epithelium can be seen through the amnion, arising from the upper and the lower explant. Arrows point to the edge of the sheet. (E) and (F) Two weeks postoperatively, the membrane fell off revealing a very clear corneal surface. Histopathology of the tissue removed confirmed presence of papilloma and intraepithelial neoplasia. (C), (D) and (E) are stained with 2% fluorescein dye. ACER, amnion-assisted conjunctival epithelium redirection.

Figure 2

Slit lamp diffuse illumination images of a patient with unilateral limbal stem cell deficiency following an alkali burn, treated with the ACER technique. (A) Preoperative image showing circumcorneal injection, corneal scarring, vascularisation and conjunctivalisation with a fibrovascular Pannus. (B) Fluorescein-stained image of the eye showing a persistent corneal epithelial defect and diffuse late staining of the epithelium, indicative of conjunctivalisation covering the entire cornea. (C) Day 5 post-ACER. Fine wrinkles are seen in the amniotic membrane (Omnigen). The cornea, visible through the amnion is very clear and transparent. (D) Fluorescein-stained image showing the amniotic membrane with some conjunctival epithelial cells on the infero-temporal margin and a large sheet of explant-derived corneal epithelium visible through the amnion. (E) The cornea appears normal after removal of the amniotic membrane 4 weeks postoperative. (F) Fluorescein-stained image of the eye showing normal corneal epithelium with no epithelial defect or late staining. ACER, amnion-assisted conjunctival epithelium redirection.

Figure 3

Slit lamp diffuse illumination images of a woman with total limbal stem cell deficiency following chemical injury with hair dye, treated with the ACER technique. (A) The eye shows 360o of fibrovascular encroachment with scarring extending up to the pupil. (B) Fluorescein-stained image showing late staining of epithelium indicative of conjunctivalisation. A small central island of normal looking epithelium is present. The vision was 0.5 logMAR. (C) Day 1 after ACER showing the amniotic membrane (Omnigen) covering the cornea and the superior explant. (D) Covering the inferior explant. (E) A shining cornea with normal transparency is seen after removal of the amnion around 3 weeks postoperatively. Some anterior stromal scarring is visible along the periphery but the vascularisation is minimal. Vision improved to 0.0 logMAR (normal). (F) Fluorescein stain shows a normal corneal epithelium with normal tear distribution across the entire cornea with no late fluorescein staining. ACER, amnion-assisted conjunctival epithelium redirection.

Preoperative IVCM of corneas and limbus showed classical features of LSCD in the form of hyper-reflective conjunctival epithelium on the cornea with goblet cells and cystic spaces (figure 4). This returned to normal corneal phenotype postoperatively in all the successful cases. Preoperative OCT showed a thickened epithelium in the peripheral cornea and variable thickness of the underlying stroma. Postoperatively, the central epithelial thickness was normal in all but two cases that failed (figure 4).

Figure 4

Representative pictures of corneal imaging with in vivo confocal microscopy (IVCM) and optical coherence tomography (OCT). (A) Preoperative IVCM of the central corneal epithelium in an eye with limbal stem cell deficiency (LSCD) showing hyperreflective conjunctival cells and clusters of goblet cells scattered individually and organised into rosettes (arrow). (B) IVCM of the limbus in a patient with LSCD showing hyperreflective conjunctival epithelial cells with cystic spaces, inter-digitating with normal peripheral corneal epithelial cells (dark cells with bright outlines). (C) Postoperative IVCM of central corneal epithelium showing normal phenotype of limbus-derived corneal epithelial cells (bar=50 µm). (D) Preoperative OCT of a patient with LSCD showing thickened epithelium (290 µm at thickest part) lying on the anterior surface of the corneal stroma delineated by a white line (arrows). The corresponding stroma measured 171 µm. The corneal thickness at the thinnest part was 110 µm. (E) Postoperative OCT showing the normal thickness of the epithelium.

Additional intraoperative and postoperative treatment

All five patients with allo-LSCT were given oral tacrolimus postoperatively and one (20%) patient required additional oral prednisolone of tapering regimen (patient 10). One patient underwent simultaneous penetrating keratoplasty and cataract surgery (patient 5) and another patient underwent simultaneous penetrating keratoplasty (patient 8) during the ACER procedure. The fibrovascular tissue excised from the cornea of patient 7 demonstrated conjunctival intraepithelial neoplasia on histopathological examination. This required additional topical interferon-alpha treatment. Complete remission was achieved following the treatment. Patient 3 (with SJS) developed a complete recurrence of LSCD after the ACER procedure and required a repeat cadaveric allo-LSCT combined with penetrating keratoplasty and cataract surgery 7 months later. Five months later, the patient developed a corneal perforation necessitating corneal glueing and was referred for consideration of osteo-odonto-keratoprosthesis (OOKP). Patient 8 (with aniridia) developed a recurrent total LSCD post-ACER and underwent a repeat cadaveric allo-LSCT with ACER 6 months later. A repeat penetrating keratoplasty and cadaveric allo-LSCT was further performed 1 year later. Unfortunately, the cornea became completely conjunctivalised and hazy 5 months later and the patient was referred for consideration of OOKP. No additional intervention such as autologous serum drops, bandage contact lens or tarsorrhaphy was performed to promote epithelialisation in any eye as the outer amniotic membrane afforded adequate cover and protection.

Discussion

In this study, we present a modified ACER technique, using vacuum-dried amniotic membrane (Omnigen) and fibrin glue, in managing total LSCD. We demonstrated that 80% of our patients achieved a complete or partial success following the modified ACER technique. The BCVA improved significantly from 1.76±0.64 logMAR preoperatively to 0.94±0.94 logMAR at the final follow-up. Among the successful patients, visual improvement was better if those with aniridia were excluded, from 1.53±0.74 preoperative to 0.23±0.25 at final follow-up (p=0.004). However, in real life, even the patients with aniridia found the improvement very useful, giving them greater independence.

We demonstrated that the modified ACER technique was effective in managing total LSCD secondary to a wide range of causes, including chemical eye injury, contact lens overuse, ocular surface neoplasia and congenital aniridia-related keratopathy. However, it is important to note that many of these conditions that are capable of damaging the limbal stem cell niche can lead to severe dry eye disease. Living tissue transplants do not survive in severe dry eye disease, hence tissue transplant procedures such as limbal or corneal grafts should not be considered in such cases. OOKP is a last-resort procedure with significant consequences when it fails. Hence, in our opinion, the modified ACER technique is always worth performing before considering OOKP. Although not used in the cases reported here, tarsorrhaphy and lubrication with autologous serum and/or artificial tears are important adjunctive considerations.

Omnigen is a dehydrated human amniotic membrane-derived matrix, manufactured using the standardised Tereo process.25 26 According to the manufacturer, Omnigen is made from fresh amnion, denuded of the spongy layer, and delicately dehydrated using a patented low temperature vacuum evaporation process, not involving freezing, lyophilisation or heat, to uniquely preserve the structure, composition and biological properties of amnion.25 The stable matrix is easy to handle and apply directly to the wound in a dry state for rapid and effective in vivo rehydration and enhanced wound healing.25 It was shown that vacuum-dried amnions (Omnigen) exhibited enhance, or at least comparable, structural properties, biochemical stability and wound healing property when compared with cryopreserved amniotic membranes.25 Moreover, Omnigen is advantaged by its ready availability off the shelf, storage at room temperature, ease of handling and importantly its transparency.

Biological properties of vacuum-dried and cryopreserved amnion are different.

The processes of handling, washing and storing of the membranes induces changes that affect the final constituents of the clinically-ready product. Studies specifically targeted towards loss of transforming growth factor (TGF)-beta and epidermal growth factor have shown a variable but significant loss during processing of cryopreserved membrane. As most of the TGF is contained in the spongy layer of the amnion, and as this is deliberately removed from Omnigen, the constituents contained in the spongy layer will be absent or markedly reduced in Omnigen. TGF promotes ‘healing’ with scarring, hence this will be less a risk with Omnigen compared with cryopreserved amnion.27 28

It must be noted, however, that both types of amnion come from human donors and there are a whole host of variations from donor to donor based on age, race, gravidity, parity, trial of labour before caesarean section, site of the amnion sac and other factors.29 30

Despite the differences and the inferred effects, it is not definitively known how much impact these have on the clinical outcomes resulting from use of the different amnions. In this study, we noted that the transparency of Omnigen allowed visualisation of the outgrowth of epithelial sheet from the explants, making it easy to monitor growth to complete healing. As reported earlier, larger sheets of epithelium, migrating from the superior explants, were seen compared with the inferior explants.31 It also provided the necessary protection to the explants and new epithelial cell from exposure and the blink action of the lids. Subjectively, as Omnigen is thinner than frozen amnion, it was easier to remove at the slit lamp by tearing or cutting with a pair of Vannas scissors, along the edge of the recessed conjunctiva.

Using fibrin glue to attach Omnigen to the sclera, and the recessed conjunctiva on the periphery of the membrane provided advantages of ease and rapidity of application, obviating suture-related potential complications and the need to remove sutures later. The fibrin clot along the edge of the recessed conjunctiva acted as a physical barrier,32 retarding the growth of conjunctival cells on to the membrane. The fibrin glue used (Tisseel, Baxter

UK) contains aprotinin, an anti-fibrinolytic agent, that prevents the clot from lysis for 10–14 days, retaining the barrier effect of the fibrin clot for that duration.32 Omnigen remained at the ocular surface for a mean duration of 3.27±1.38 weeks. This was not statistically different from the duration for which cryopreserved amnion stayed on the ocular surface (2.10±1.50 weeks).10 Cost-effectiveness of vacuum-dried amnion (Omnigen) and cryopreserved amnion has not been evaluated. However, as Omnigen demonstrates similar effectiveness to cryopreserved amnion without the need for expensive facility for cryopreservation, Omnigen is likely to be at least as cost-effective as the cryopreserved amnions. Future studies examining this important aspect would be of value. The use of oral tacrolimus instead of high dose oral prednisolone following allo-LSCT was based on our previous clinical experience where we demonstrated tacrolimus as a useful treatment in preventing/reducing the risk of corneal and limbal graft rejection.22 23 Three (75%) of the four patients that were managed with oral tacrolimus postoperatively (without oral prednisolone) did not experience any limbal graft rejection, underlining the efficacy of this regime. However, high-dose oral prednisolone is recommended in the case of high-risk cases, especially in eyes that had a previous history of limbal graft rejection.

SLET and Holoclar are in vivo and ex vivo expansion techniques respectively, designed to achieve the same outcome. In SLET, the epithelial cells from the tiny pieces of donor limbus tissue attached to the cornea proliferate and migrate to cover the entire cornea. There is no attempt made, in the surgical design, to prevent encroachment of the conjunctival epithelium. As small pieces of limbal tissues are transplanted over the entire denuded cornea, simultaneous re-epithelialisation of the cornea from the donor-derived epithelium across the entire cornea may prevent conjunctivalisation of the temporal and nasal aspects of cornea, which are commonly the last areas to heal.33 However, in a large study of 125 eyes with chemical eye injury and unilateral LSCD that underwent SLET, it was noted that 76% eyes achieved successful outcome at 1 year, with 18.4% eyes suffering from progressive conjunctivalisation.34 When a small piece of limbal tissue is cut into several tiny bits, it is acknowledged that some of these would have epithelium and some would be stromal tissue only. Not all bits affixed to the corneal surface (covered with amnion) will yield a growth of epithelium, leaving gaps until confluence is attained. Moreover, the bits are placed in the mid periphery to two thirds periphery, which would enable conjunctival epithelial encroachment across the limbus to cover the peripheral cornea. Similarly, Holoclar provides an ex vivo expanded sheet of limbal epithelial cells to cover the cornea but it too does not address the issue of any conjunctival cells encroachment on the cornea. However, both SLET and Holoclar give successful outcomes.

This study was limited by the small sample size, with heterogeneous causes of LSCD. However, we demonstrated that 80% of the patients achieved complete/partial success following LSCT, highlighting that this procedure is versatile and beneficial for various indications. In addition, this was a non-comparative study specifically examining the effectiveness and safety of ACER technique using vacuum-dried amnion (Omnigen) only. Future comparative studies, ideally randomised controlled trials, comparing the effectiveness and safety of ACER technique between vacuum-dried amnion (Omnigen) and cryopreserved amnion will help further inform the clinical practice.

Two unresolved questions remain. (1) How does the corneal epithelium survive ‘indefinitely’ without restoration of the limbal niche? The answer to this could be related to the stem cell potential of the transient amplifying cells (basal epithelial cells) as previously reported.21 Central islands of epithelium have been shown to survive in corneas with total LSCD.21 35 Another plausible answer could be related to the evidence provided by some but not all studies, that donor-derived DNA is absent in the recipient corneas within 20 weeks after limbal stem cell transplantation, suggesting that once the normal environment is re-established, bone marrow-derived stem cells of the recipient could home to the cornea and ‘transdifferentiate’ to the corneal epithelial phenotype.36–38 (2) What will happen to the healing response if an injury were to denude a large central area of the cornea? Anecdotal evidence suggests that the healing response is poor and conjunctivalisation follows.

Data availability statement

Data are available upon reasonable request. Individual patient data are stored in the hospital’s electronic system and can be retrieved on reasonable request.

Ethics statements

Patient consent for publication

References

Footnotes

  • 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 HSD is consultant to Artic Vision, Electrospinning, Santen, Thea and has shares in Glaxosmithkline and NuVision biotherapies (manufacturers of Omnigen) and is joint patent holder for the vacuum drying method for preparing Omnigen). None of the other authors have any conflict of interest to declare. This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors. HSD has received an unconditional educational grant from Santen, which is not related to the content of this paper.

  • Provenance and peer review Not commissioned; externally peer reviewed.