Mechanical Debridement

If the visual axis or a larger portion of the peripheral cornea is covered by conjunctival tissue, simple mechanical debridement of this tissue may allow the remaining corneal stem cells to repopulate the central cornea with normal or near-normal epithelium. The procedure can be done with topical anesthesia and consists of debriding abnormal epithelium with a crescent blade or Weck–Cel sponge followed by the application of a bandage contact lens. The goal of this procedure is to provide the patient with a fairly clear visual axis and not to make the entire corneal surface normal. Some investigators have reported success with as little as two clock hours of normal limbal cells. In a variant known as sequential sector conjunctival epitheliectomy (SSCE), conjunctival sheets are debrided every 24–48 hours until the patient has completely re-epithelized. Regardless of the number or extent of debridement, patients should be placed on postoperative topical antibiotics and artificial tears. This procedure can be done in combination with an amniotic membrane graft as described below.

Amniotic Membrane Grafting

The most important properties of amniotic membrane harvested from the innermost layer of the placenta include the anti-inflammatory effect through downregulation of fibroblasts and providing a substrate for proliferation of the corneal and conjunctival epithelial cells. These properties are only useful when some reserve of stem cells is present because the amniotic membrane itself is not a source of stem cells.

Work by Tseng and others has shown that the application of amniotic membrane to eyes with partial stem cell deficiency can improve ocular surface health and, in some cases, even restore a near-normal corneal epithelium. Examples include acute and chronic chemical injuries, acute Stevens–Johnson syndrome, and iatrogenic stem cell deficiency. In these cases, application of the amniotic membrane, with or without mechanical debridement, can result in decreased ocular inflammation and allow the remaining corneal stem cells to repopulate the ocular surface.

Amniotic membrane is commercially available in several forms, including a preserved wet amniotic membrane, a dehydrated form, and a contact lens–mounted membrane. The amniotic membrane is placed on the eye with the epithelial side up to cover the area of interest. The membrane then can be secured to the cornea or episclera with 10-0 nylon sutures or fibrin glue ( Figs. 4.30.2 and 4.30.3 ).

An amniotic membrane application shortly after severe chemical burn.

The same patient as in Fig. 4.30.2 , 2 years after the placement of amniotic membrane tissue.

Autologous Limbal Stem Cell Transplantation

When there is relative or sectoral stem cell deficiency and the condition is unilateral, the unaffected part of the eye or the contralateral eye can serve as a donor for stem cells. Stem cells can be harvested from either the contralateral eye or from normal areas of the affected eye and transplanted to the area of stem cell deficiency. The first option is preferred. The surgical technique used is identical to that for an autologous graft for total stem cell deficiency and is described below.

Autologous Limbal Stem Cell Transplantation

The traditional surgical approach involves careful dissection of the conjunctival tissue from the limbal area of the affected eye. A rim of keratolimbal tissue from the contralateral eye, including a conjunctival rim and anterior cornea, can be harvested. Several tissue grafts can be harvested from the fellow eye. The superior and the inferior 4–5 clock hours of limbus offer the highest concentration of corneal stem cells. This tissue then can be carefully transported to the prepared bed with the maintenance of correct orientation. The tissue can be sewn into place by using nonabsorbable sutures, taking care not to pass the sutures through the harvested stem cell area. However, now that fibrin glue is widely available, it can be used to attach this tissue, thus avoiding the inflammatory reaction associated with sutures. A bandage contact lens is then applied to the eye.

More recently, Sangwan et al. described SLET, a novel technique. In this procedure a 2 × 2 mm or 1 clock hour section of autologous limbal tissue was excised from the uninvolved eye. The harvested graft was then divided into eight segments. After the host cornea was scraped of any fibrovascular epithelium, and an amniotic membrane secured with fibrin glue and the segments were placed epithelial side up on the optimized ocular surface and secured with fibrin glue, a bandage contact lens was then fitted to protect the tissue.

Postoperatively, patients are treated with antibiotics, corticosteroids, and preservative-free artificial tears. The bandage contact lens can be removed when the ocular surface is stable. Long-term systemic immune suppression is not necessary in these patients.

Prior to undertaking the harvesting of limbal tissue for stem cell transplantation, it should be clearly established that the fellow eye is not affected, and that there is an ample reservoir of stem cells present. A risk of inducing relative stem cell deficiency to the donor eye does exist if too much tissue is harvested. Most clinicians avoid taking more than 4 clock hours of tissue from the donor eye. When concern exists about the health of the fellow eye, allografts may be considered (see below). This method is highly successful, with graft survival rates reported up to 100% over a 47-month period.

Basu et al. recently reported long-term follow up of 125 cases and found that with this approach, patients had notable improvement in the clinical appearance of their ocular surface as well as an increase in their visual acuity during the 1.5-year follow-up. The clinical factors associated with failure were identified as acid injury, severe symblepharon, and SLET combined with keratoplasty, In addition, the chapter authors’ group has observed poorer outcomes in SLET patients who had autoimmune cicatrizing conjunctivitis as the cause of their LSCD. Another indication for SLET is LSCD after treatment of ocular surface neoplasia. It should be noted that in this scenario, it would be prudent to defer ocular surface reconstruction of the affected eye until it has been established that there is no recurrence of the neoplasia.

Keratolimbal Allograft and Allogeneic SLET

Keratolimbal tissue can be harvested from either a living-related donor’s healthy cornea or from the whole globe or from corneal tissue of a cadaveric donor. The harvested allograft can then be transplanted directly onto the injured eye. Neither living-related donors nor cadavers are perfect sources for grafts, and thus both advantages and disadvantages must be weighed.

Living-related donor transplants are advantageous in that harvested cells are typically more immunologically compatible than from a random source, especially if human lymphocyte antigen (HLA) matching is performed. Reinhard et al. demonstrated that this allows for longer survival time of the transplanted cells. Specifically, they showed that over 5 years the grafts with only 0–1 HLA mismatches had a 65% success rate compared with 14% of unmatched tissue. Living-related donor tissue also has the benefit of stem cells that are fresh. One recent modification is the use of allogeneic SLET, which is surgically identical to SLET described above except that the donor eye is a living-related donor or cadaver rather than the patient’s fellow eye. The disadvantage of all allogeneic limbal stem cell techniques is that even with HLA matching, postoperative systemic immunosuppression is required. Furthermore, the amount of tissue that can be harvested from the donor eye is limited, and there is a risk of inducing an ocular surface disorder in the healthy eye of a donor.

For keratolimbal allograft (KLAL), the tissue can be harvested under peribulbar anesthesia, similar to what has been described above for conjunctival limbal autograft. A total of 4 clock hours of tissue may be harvested, equally divided from the upper and lower limbus. This donor tissue then is placed onto the recipient eye after peritomy and superficial keratectomy have been performed to prepare the limbus. The donor tissue is secured to the cornea anteriorly and posteriorly to the episclera using either fibrin glue or 10-0 nylon sutures. Amniotic membrane may be transplanted at the same time, as needed.

In contrast, using cadaveric donor tissue allows the surgeon to harvest greater quantities of tissue than can be removed from living eyes. This theoretically improves the success of the ocular surface reconstruction, especially in eyes with severe stem cell deficiency. However, cadaveric tissue must be screened for communicable diseases. The time necessary for these tests along with tissue damage that occurs during standard tissue processing can result in stem cells of lower quality. The Minnesota Lions Eye Bank suggested certain criteria for tissue harvested for KLAL. These criteria include obtaining tissue from the youngest donor possible (including pediatric sources), taking care to avoid damage to limbal stem cells, including a 3–4 mm skirt of conjunctiva and a large corneoscleral rim, as well as obtaining both donor eyes to ensure adequate tissue.

Some investigators advocate the use of more than one donor eye to allow complete coverage of the limbus. Penetrating keratoplasty (PKP) or anterior lamellar keratoplasty (ALK) can be performed at the same time, or at a later setting when the ocular surface is more stable. In 2010, Choi et al. reported a modified technique for both graft harvesting and host bed preparation by using a femtosecond laser, which showed good short-term results.

Postoperatively, patients are treated with topical antibiotics, corticosteroids, and systemic immunosuppression. Because of the relative abundance of Langerhans’ cells and HLA-DR antigens in the limbus, a high rate of immunological reaction can be expected, which may lead to recurrence of stem cell failure, necessitating aggressive and long-term immunosuppression. Although experts agree that postoperative long-term immunosuppressive therapy is necessary for the success of nonautologous grafts, the exact regimen and timeframe for treatment are highly variable. One commonly used method for postoperative immunosuppression is the Cincinnati protocol. This technique employs 1 year of corticosteroid coverage along with systemic tacrolimus, sirolimus, and mycophenolate. It is not yet clear to what extent and for how long patients require immunosuppression after living-related and cadaveric allogeneic SLET.

The reported outcomes of KLAL vary, depending on the underlying disease entities and the duration of follow-up. A recent study reported successful improvement in ocular surface in up to 89% of living-related donors over a 32-month follow-up. In contrast, the same study reported only a 33% success rate for cadaveric donor grafts over the same period. However, the long-term follow-up has shown a trend toward progressive decline of stem cell population and destabilization of the ocular surface despite systemic immunosuppression.

Ex Vivo Expanded Limbal Stem Cells and Nonocular Tissue

The majority of recent advancements in the area of stem cell transplants have focused on the harvesting and subsequent ex vivo expansion of limbal cells. Expansion of stem cells by culturing them in vitro theoretically provides a large supply of stem cells that can be used for surface reconstruction. This allows the surgeon to not only selectively remove fibroblast and Langerhans’ cells, which may affect the long-term survival of allografted cells, but also allows for minimal tissue to be excised from a donor source, decreased the potential risks associated with tissue harvesting.

In this approach, a minimal amount of limbal tissue (1–2 mm) is harvested either from an eye with relative stem cell deficiency, or the normal contralateral eye of a patient with unilateral total stem cell deficiency. In cases where severe and total bilateral stem cell deficiency exists, cells can be harvested from either a living-related donor or cadaver eyes. Studies have demonstrated the ability to use conjunctival tissue as well. These cells then are amplified in culture media on a carrier, which will be utilized for transport and transplant of the cells onto the diseased eye. The exact method of ex vivo expansion varies widely, with success demonstrated under many conditions, including both explants and cell suspensions with or without 3T3 mouse fibroblasts and serum. However, regardless of the method for culturing the cells, obtaining a high percentage of p63-bright cells (>3% of all clonogenic cells) is of vital importance to the success of the graft.

The amplified cells then may be mounted on a substrate. Current substrates include petrolatum gauze, denuded human amniotic membranes, fibrin, 3T3 cells, and bandage soft contact lenses. The recipient eye is then prepared in a similar manner to the method described for keratolimbal grafting. The cultured stem cells and their carrier are transferred onto the recipient bed, anchored to the limbus with 10-0 nylon sutures and to the surrounding conjunctiva with 8-0 Vicryl sutures. A bandage contact lens often is placed on the eye and kept in place until the ocular surface stabilizes.

Although postoperative treatment for allogeneic tissue is similar to that for patients with keratolimbal grafts as described above, autologous tissue should be used when possible to avoid the need for immunosuppression.

Schwab et al. found improvement in the ocular surface of 60% of patients with autologous cells, and in all of the patients (total 4) with allogeneic cells combined with immunosuppression, with a mean follow-up period of 13 months. Shimazaki et al., on the other hand, found only a 46.2% success rate in achieving a stable and healthy ocular surface in allografted patients. Furthermore, in this report, the authors did not find a difference in success rate between this technique and cadaveric limbal transplantation combined with amniotic membrane. Baylis et al. recently reviewed 28 case reports and series regarding cultured limbal stem cells published over a 13-year period and compiled outcome data. Despite wide variation in technique, they noted an overall success rate of 77% for autografts and 73% for allografts (76% overall). They also demonstrated that failures typically occurred in the first 2 years before stabilizing. In a 10-year study of 113 eyes reported in 2010, Rama et al. showed a 76% success rate, with most failures occurring in the first year. They noted that, as mentioned earlier, grafts with more than 3% p63-bright cells had a 78% chance of success versus only 11% in those grafts with fewer than 3%.