To our knowledge, transplantation of posterior corneal tissue has first been performed clinically by Barraquer, who made a microkeratome anterior lamellar flap and replaced a posterior lamellar disc underneath the flap. In 1993, we started with a different approach for posterior lamellar keratoplasty, in which the anterior cornea was left intact, and a posterior lamellar disc was transplanted through a scleral incision and deep stromal pocket.

Technique A

In animal studies, a 6.0 mm posterior lamellar disc was transplanted through a 9.0 mm scleral incision, after making a dissection at 50% corneal depth in both donor and recipient tissues (Technique A, see chapter 4). To perform the procedure clinically, the technique was slightly modified. Instead of a 6.0 mm diameter posterior lamellar transplant used in the preceding animal study, the patients had a 7.0 or 7.5 mm diameter transplant. Instead of using a keratotomy knife to obtain the appropriate stromal dissection depth, the anterior chamber was filled with air to visualize the posterior corneal surface, so that the depth of dissection relative to the corneal thickness could be monitored optically (See chapter 2). Furthermore, a stromal pocket was dissected at 80% instead of 50% corneal depth. With deeper dissections, the excision of posterior corneal tissue may be easier, less interface scarring may be induced, and a thicker portion of the anterior stroma is unaffected, so that postoperative astigmatism may be further minimized.

In a series of 20 patients, all posterior transplants cleared and kept their position without suture fixation. Although the anterior chamber was completely filled with air at the end of the surgery, an air bubble may not be neccessary to fixate the posterior transplant. In our second patient, who was made aphakic during the keratoplasty procedure, the air bubble in the anterior chamber escaped through the pupil into the vitreous cavity within the first hour after surgery. Still, loosening or dislocation of the transplant did not occur. It may therefore be speculated that the stickiness of the stromal tissue at the donor-to-host interface, the suction force of the donor endothelium, fibrin deposition throughout the wound area in the early healing phase, and stromal repair in later phases, keep the posterior transplant in position.

Except in the first two patients who had pre-existing maculopathies, a useful spectacle corrected visual acuity was generally achieved at six months after surgery. All patients showed relatively low with the rule astigmatism. The suture induced astigmatism may be reduced by removing the sutures from the scleral incision at three months after surgery. However, because the astigmatic error could be easily corrected with spectacles in all patients, scleral suture removal was not performed.

Figure 9.1: The first patient who had a 7.0 mm posterior lamellar transplant, at two year after surgery. At 12 oclock, the 9.0 mm scleral incision is seen through which the donor tissue was implanted. The iris claw intraocular lens was implanted following a complicated cataract extraction in the 1980s.

Since donor posterior corneal tissue was harvested during surgery from whole human eye bank eyes, preoperative endothelial cell counts could not be obtained. An average postoperative endothelial cell density of about 2500 cells/mm2 measured at an average time interval of nine months after surgery, may indicate that sufficient donor endothelium was transplanted (and had survived). This agrees with our findings in a preceeding eye bank eye study, in which posterior transplants showed about 1% of dispersed, focal cell death, and a zone of endothelial cell loss at the outer edge due to trephination of the tissue. After posterior lamellar keratoplasty endothelial cell loss may be expected to occur from both the surgical trauma with subsequent endothelial migration underneath the wound area, and the normal rate of endothelial cell loss after transplantation. In addition, the presence of an air bubble in the anterior chamber during the first postoperative days may have caused endothelial cell damage.

The first two cases showed postoperative pachymetry readings of less than 500 um, following stromal dissection of the donor and recipient cornea at 80% depth. Thinning of the donor posterior tissue after surgery may be explained by the fact that relatively more post mortem hydration, i.e. swelling is present in the posterior than anterior cornea. Thus, the actual depth of dissection in a post mortem cornea may be deeper than intended. For example, a dissection made at 60% depth in a post mortem, swollen cornea may equal an 80% dissection depth in a physiologically hydrated cornea. Thus, in donor corneas used for the other patients, the dissection was made at 60% depth, to obtain an eventual recipient corneal thickness of approximately 550 um.

All patients showed a normal healing response after surgery. Although excessive scarring at the donor-to-host stromal interface did not occur, haze formation was assumed to reduce the BCVA in five patients (See chapter 3). For this reason, one patient had a penetrating keratoplasty at one year after surgery. With light microscopy, this human specimen, as well as the monkey specimens, showed a normal fibroblastic response, but distortion of the collagen fibers in the scar area. If effective healing occurs at the unsutured, posterior transplant wound, long-term fixation of the donor tissue may be achieved, i.e. late dislocation of the transplant may seem unlikely. Formation of a retrocorneal membrane did not occur.

Technique B

Disadvantages of Technique A included the large scleral incision inducing "with the rule" astigmatism, and the use of the intracorneal trephine and "spoon shaped glide" obscuring the view of the anterior chamber structures during trephination and implantation of the donor. Furthermore, the technique was limited to transplanting a posterior lamellar disc with a maximum diameter of 7.5 to 8.0 mm. Therefore, the technique was further modified to transplant an 8.5 to 9.0 mm donor disc through a self sealing scleral tunnel incision (Technique B, See chapter 5), without the use of any (corneal or scleral) sutures.

Figure 9.2: The arrows point to a 8.5 mm posterior lamellar donor disc, that was transplanted through a self-sealing, 5.0 mm scleral tunnel incision (arrowheads).

With experience, a dissection depth of 95% to 98% corneal depth can be obtained, so that the remaining posterior corneal layers are relatively thin, and more easily excised. Again, the more anterior corneal tissue is left intact, the less astigmatism may be of induced by the surgery.

In July 2000, the first patient had a posterior lamellar keratoplasty through a self-sealing incision. One week after surgery, the best corrected visual acuity was 0.8. Since then, the technique has been performed in another six patients. Although the number of patient is too small to reach any conclucions, Technique B appears clearly superior to Technique A (See chapter 5).

In conclusion, our preliminary clinical results suggest that posterior lamellar keratoplasty through a 9.0 mm sclero-corneal pocket incision, or a 5.0 mm scleral tunnel incision, is a feasible surgical approach to manage corneal endothelial disorders. Compared to penetrating keratoplasty, posterior lamellar keratoplasty may have the advantages of less postoperative astigmatism and faster visual recovery, a minimal risk of late wound dehiscence, and elimination of all suture related complications.