Statistics from Altmetric.com
Editor,—While we agree with Dua and Azuara-Blanco1 that the use of “fresh” donor eyes for limbal transplantation is preferable, social and surgery scheduling limitations may force surgeons to use stored corneoscleral (C-S) rims. Tsubota et al 2 have demonstrated the viability of limbal stem cells (SCs) harvested from C-S rims stored in Optisol GS for up to 5 days. The surgical method described by Dua and Azuara-Blanco1 positions the limbal allograft posterior to the anatomical limbus in the host. The 150 μm thick donor limbal graft can result in a stepped ocular surface, which can be detrimental to long term survival of the transplanted epithelium. Splitting the ring of limbal tissue and interposing a separate piece of corneal stroma or limbus theoretically allows chinks in the reconstituted limbal barrier. Finally, the technique described does not allow sufficient flexibility in titrating the thickness of the donor tissue used. Since an important goal of ocular surface reconstruction is to achieve a smooth surface, surgeons often have to use donor limbal grafts of differing thickness in individual recipient eyes. This flexibility is possible if the limbal graft is fashioned from a C-S rim. In eyes undergoing combined penetrating keratoplasty and limbal transplantation, apposing the donor limbal and corneal graft without an intervening gap is preferable.
The surgical failure in case 4, who underwent limbal allotransplantation 3 weeks after severe alkali burns, corroborates our recent report on the optimal timing of limbal transplantation after ocular surface burns.3 We feel that complex procedures such as limbal autografting or allografting are best performed after resolution of ocular surface inflammation and limbal revascularisation. While use of autoserum tears is beneficial, ostensibly by providing biological factors promoting epithelial health, corneal immunoglobulin deposition has been reported in a patient with persistent epithelial defect.4
Four of the eyes in this report have a follow up of less than 1 year. Both eyes with follow up greater than 1 year experienced a graft rejection episode—during FK-506 therapy in one eye and after cessation of the drug in the other. This emphasises our limited understanding of the immunology of this procedure. Despite the initial encouraging report by Tan et al,5 our experience in a larger cohort with longer follow up, indicates that HLA matched limbal transplants from live related donors have poor long term survival, in the absence of systemic immunosuppression.6We thus, agree with Dua and Azuara-Blanco1 that cadaver limbal transplantation offers the advantages of greater limbal stem cell transfer and is probably the procedure of choice despite the need for systemic immunosuppression of the recipient. A controlled clinical trial is difficult in this condition and reports like that of Dua and Azuara-Blanco are required to improve our understanding. We congratulate them on their use of FK-506 in limbal allografting.
Editor,—Rao et al have raised several issues with regard to limbal stem cell transplantation. The use of fresh versus preserved donor tissue remains unresolved. Empirically it is generally considered that fresh limbal tissue is better than stored. Storage conditions vary in different countries. In the UK for instance, by far the largest supply of donor material is stored in organ culture medium (Eagle's MEM) with dextran added to deturgesce the tissue before use. This material has proved to be excellent for corneal transplantation (up to 4 weeks in storage) but has not been used for allo-limbal transplantation. Although corneal epithelial cell cultures can be established from such donor rims, its efficacy as a source of limbal stem cells remains to be tested.
The thickness of 150 μm includes 50 μm or more of limbal epithelium. The thickness of stromal tissue is therefore less than 100 μm. This is largely to facilitate handling and suturing of tissue. In our experience, the development of a “stepped ocular surface” was not an issue. In fact, over a period of several months, the tissue thinned and merged imperceptibly with the host. The “long term survival” of the epithelium was never compromised by the thickness of the donor limbus. The titration of donor tissue thickness is only relevant if a recipient bed is being fashioned to receive the donor tissue. This is often the case in auto-limbal transplantation. Placement of the donor limbus posterior to the “perceived” anatomical limbus of the host (often it is not possible to absolutely certain where the original limbus of the host is), has the advantage of allowing use of a wider limbal rim, to include limbus and peripheral cornea. “Transient cells” have been shown to be present in the peripheral cornea. Posterior placement also makes it technically easier to perform a corneal graft should one be required at the time.
The risk of introducing “chinks in the limbal barrier”, allowing ingress of conjunctival epithelium, is only theoretical as the authors themselves have stated. The use of a “spacer” or an extra bit of limbus from the other donor eye has proved to be quite successful. Even if a complete donor limbal ring is used, it is important to watch the healing conjunctival epithelium from the recessed conjunctiva. At times, although the ring may be complete, the overlying epithelium may be missing in sectors. Conjunctival epithelium can cross over such a defect and encroach on to corneal surface. In such a situation, the principles laid down by Dua should be employed in the management.
We agree with the authors that the chances of failure are high when allo-limbal transplantation is undertaken during the acute stages of a chemical insult. This is particularly relevant when living related donor tissue is available. Such material must be reserved for use until after the acute inflammatory process has subsided. If limbal transplantation is considered essential in the early stages, serious consideration must be given to use of cadaveric donor tissue only.
Our experience with tacrolimus (FK-506, Prograf, Fujisawa Ltd, London) has been very good thus far. Several of our patients have been followed for over a year now. Attempts to reduce FK-506 (with a view to stopping treatment) have resulted in rejection reactions in two patients (one more since publication of the paper), but resolved on increasing the dose. In one patient where a rejection reaction was observed 4 months after transplantation, it corresponded to a very low trough level of the drug. We have also used this drug in the treatment of uveitis and several “high risk” corneal transplants with excellent results (unpublished observations). There is no doubt, like the authors have mentioned, that the therapy has to be continued long term.