Approximately 38,000 corneal transplants are performed in the United States each year, and the most common reason for transplantation is corneal endothelial failure. 1–3 The current standard surgical procedure for replacement of the endothelium is full-thickness penetrating keratoplasty. Although penetrating keratoplasty has a high success rate for clear grafts, the visual outcome is often delayed by irregular astigmatism, and the final visual outcome may not be stabilized until sutures are removed even years after the initial surgery. 4,5 Suture complications, such as infectious keratitis, corneal ulceration and melting, and suture-initiated graft rejection, compound the risk to the transplant that is successful in its sole objective of transplanting a healthy endothelium. 6,7 Finally, in combined procedures of cataract surgery and simultaneous corneal transplantation, the intraocular lens power selected depends on the anticipated final central keratometry power of the grafted corneal tissue. This estimated final corneal topography in penetrating keratoplasty is at best an educated guess with a wide range of mismatch between the resultant corneal power and intraocular lens power, often resulting in a myopic or hyperopic final refractive error for the patient.
The surgery of lamellar keratoplasty is based on the principal of replacing only the diseased portion of the cornea and leaving the remainder intact. Until recently, lamellar keratoplasty was possible for only diseased anterior tissue replacement, in corneas with healthy endothelium; these lamellar corneal grafts suffered from the same suture and topography uncertainties as did the penetrating keratoplasty. 8,9 Melles et al. 10–12 have recently designed a new surgical technique for posterior lamellar keratoplasty that allows replacement of the endothelium through a limbal incision with a deep lamellar plane. Their technique obviates the need for surface corneal incisions or sutures and thus has the potential of virtually eliminating the cardinal problems associated with modern penetrating keratoplasty.
We evaluated the topography of the Melles method of deep lamellar endothelial keratoplasty. We redesigned the instrumentation, slightly modified the procedure, and then analyzed the topography immediately after deep lamellar endothelial keratoplasty in eye bank eyes. Specifically, we tried to quantify the amount of change in corneal astigmatism and the change in total corneal power resulting from this procedure. We also evaluated the stability of the transplanted posterior disc of tissue that is held in place without sutures in the eye bank model in which endothelial pump function was absent or minimal.
MATERIALS AND METHODS
Sixteen human cadaver eyes from 14 donors were used in this study. These eyes were not suitable for clinical use. All eyes were processed by our osmotic thinning method previously described 13 until the corneas achieved normal physiologic thickness and clarity. Eight eyes served as recipients. Eight eyes had the corneal scleral cap excised and placed in Optisol-GS (Chiron, Irvine, CA, U.S.A.) and then were stored at 4°C until use. In no case was the recipient eye and the donor tissue from the same donor.
Recipient eyes were placed in an eye holder and secured with pins through the optic nerve. A 21-gauge needle was placed through the optic nerve into the vitreous cavity and attached to a saline gravity infusion line to maintain normal pressure. The limbus at the 12 o'clock position was marked for orientation with a marking pen. Each recipient globe underwent Orbscan analysis (Orbtek, Salt Lake City, UT, U.S.A.) of surface topography before and immediately after surgery.
Recipient Operative Procedure
A limbal scleral incision, with a chord length of 9.0 mm, was made by a trifaceted guarded diamond knife approximately 1.0 mm peripheral and concentric to the superior limbus. The 12 o'clock orientation mark bisected the incision. The knife was set for a depth of 350 μm. An angled crescent tunnel blade (Wilson Ophthalmics, Mustang, OK, U.S.A.) was used to begin the deep lamellar sharp dissection from the incision into clear cornea, extending approximately 3.0 mm centrally into the peripheral cornea. A special semi-sharp dissector spatula (Bausch and Lomb, St. Louis, MO) (Fig. 1) was then used to extend the lamellar pocket over the entire corneal diameter in every direction. A separate 0.5-mm limbal stab incision into the anterior chamber was then made with a Super Sharp blade (Katena, Denville, NJ, U.S.A.) at the 2 o'clock position of the limbal region, and air was injected by a 30-gauge cannula into the anterior chamber, filling the chamber with air and maintaining the air bubble at all times. A specially designed low-profile intralamellar trephine (Terry Trephine, Bausch and Lomb, St. Louis, MO, U.S.A.) with a cutting diameter of 7.0 mm and an external knurled circular handle (Fig. 2) was then used to trephinate the posterior lamellar recipient tissue until the anterior chamber was entered and air escaped through the deep lamellar plane. Intralamellar corneal scissors were then used to cut the remaining stromal fibers of the trephine cut and the recipient posterior stromal disc was then grasped with angled Kelman forceps (Storz, St. Louis, MO, U.S.A.) and removed from the eye through the lamellar pocket wound. The resultant recipient bed diameter was measured with external calipers and, because of the final scissors excision, varied between 7.0 and 8.0 mm. The residual central anterior corneal thickness of the recipient cornea was not measured directly, but based on several measurements of resected recipient discs, the estimated residual thickness of the anterior corneal stroma was between 300 and 450 μm. Air was repeatedly injected through the side port incision to maintain the anterior chamber during the entire excision process.
Donor Operative Procedure
The donor corneoscleral cap tissue was removed from the Optisol solution and mounted into a modified artificial anterior chamber device (Ophthalmic Specialties, San Gabriel, CA, U.S.A.) with sodium hyaluronate (Healon, Pharmacia and Upjohn, Kalamazoo, MI, U.S.A.) placed on the endothelial surface. The artificial anterior chamber was maintained with Healon and Optisol-GS solution, and with the tissue secured, the intraocular pressure was increased to higher than 30 mm Hg. The diamond knife, crescent blade, and lamellar dissection spatula were then used to create the deep lamellar pocket plane across the entire corneal diameter. The donor corneoscleral cap tissue was then removed from the anterior chamber and placed on a Brightbill Teflon block (Storz) with the epithelial side down. A disposable trephine (Katena) was then used to cut the central donor disc by cutting from the endothelial side through to the epithelial side. The diameter of the donor trephine varied between 7.0 and 8.0 mm, and the diameter was chosen to match the measured diameter of the recipient bed perfectly. The thickness of the donor button was not measured directly to avoid distortion of the tissue before transfer. The estimated thickness was between 200 and 300 μm. Two fine forceps were then used to grasp the stromal edges of the trephine cut, and the deep posterior stromal disc of tissue with the donor endothelium was separated from the remainder of the donor corneoscleral cap. This endothelial donor disc was then gently transferred and placed, endothelial side down, onto a specially designed, Healon-coated, transfer spatula (Bausch and Lomb) (Fig. 1).
Transplantation of Tissue
The anterior chamber of the recipient eye was then fully inflated with an air bubble. The transfer spatula with the donor tissue was then entered into the pocket wound, dropped down into the anterior chamber, and then lifted gently upward placing the donor disc of tissue into the recipient bed of the resected central area and placing the donor stromal side against the recipient stroma. The transfer spatula was then gently retracted from the eye, leaving the donor tissue disc self-adhering to the central posterior recipient bed. Minor adjustments to the coaptation of the edges of the donor disc and recipient bed were made with a Sinskey hook (Stephens Instruments, Lexington, KY, U.S.A.) through the side port incision. The superior pocket wound was then closed with three or four interrupted 10-0 nylon radial sutures, and the air bubble was replaced with balanced salt solution through the side port stab incision. No adjustment of sutures was done to manipulate the astigmatism. Stability of the disc was tested by manually shaking the recipient globe and pounding on the central cornea and limbal surfaces with not only the tips of a smooth blunt forceps but also the handle of the forceps.
Each recipient eye then underwent Orbscan analysis for postoperative corneal topography. Specifically, the net change in astigmatism and the vector-corrected change in astigmatism before and after surgery and the change in net central corneal dioptric power before and after surgery were evaluated. Statistical analysis was by Student t test. Vector analysis was performed using the method of Jaffe and Clayman. 26
Each of the eight recipient eyes was analyzed by Orbscan for preoperative and postoperative astigmatism and also central corneal diopter power. The results for each eye are shown in Tables 1 and 2. The mean change in net corneal astigmatism was 0.4 ± 0.5 diopters, showing no significant difference between preoperative and postoperative corneal astigmatism (p = 0.22). The axis of the steep meridian showed an insignificant shift from a mean axis of 103° to 115° (p = 0.59). The mean net change in corneal power was 0.2 ± 0.4 diopters of corneal flattening, showing an insignificant change in corneal power from before to after surgery (p = 0.27). A qualitative topographic map comparison between the preoperative and postoperative maps of eye 4 can be seen in Figure 3. The size, shape, and power of the central astigmatic area changed little after the deep lamellar endothelial keratoplasty procedure.
External attempts at dislocation of the donor disc with blunt trauma and also by vigorous shaking of the globe were unsuccessful. The disc was displaced only by using a Sinskey hook from within the anterior chamber, by peeling the edge of the disc from the recipient bed edge. The stability of the disc in the recipient bed was remarkable.
Endothelial cell decompensation from conditions such as Fuchs' endothelial dystrophy, pseudophakic or aphakic bullous keratopathy, and viral keratitis can lead to visual loss from corneal edema and represents the primary reason for penetrating keratoplasty. 1–3 Although modern technology, techniques, and eye banking have yielded an astonishingly high success rate of clear grafts and functioning donor endothelium, visual recovery is relatively slowed by donor epithelial sloughing, irregular astigmatism, and a host of suture-related problems.
The most vexing problem for the corneal surgeon is the unpredictability of corneal topography after penetrating keratoplasty. Various suture techniques with intraoperative adjustment and postoperative selective suture adjustment or removal have reduced, but not eliminated, the incidence of exceedingly high astigmatism. 4,5,14 Nonetheless, early postoperative astigmatism levels vary between 3 and 7 diopters and are rarely less than the baseline preoperative astigmatism in patients without keratoconus. 1,4,5 In my own hands (M.A.T.), early suture-in astigmatism after penetrating keratoplasty has averaged 2.42 ± 1.84 diopters in patients with Fuchs dystrophy and averaged 2.13 ± 1.64 diopters in patients with keratoconus when using a single continuous running suture technique with meticulous intraoperative suture adjustment and no postoperative suture adjustments. 15 These results are similar to those reported by Serdarevic et al. 25 using the same technique. The extensive wound healing that must take place in penetrating keratoplasty limits our capacity for control of final topography once all sutures are out, and all experienced corneal surgeons have experienced the disappointment of a clear corneal graft that unpredictably shifts from low to high astigmatism when the final suture breaks or is removed, even years after the initial surgery. A procedure that would allow replacement of the diseased host endothelium without violating the surface epithelium or anterior corneal tissue with incisions or sutures would solve many of our early postoperative problems and also require less extensive wound healing to influence long-term astigmatic results.
The inherent philosophy of lamellar keratoplasty is to replace only the diseased tissue and leave the remaining cornea intact. This approach has been extensively applied in the treatment of anterior corneal disease 16–18 but only recently has been attempted for endothelial decompensation. 10,12,19 Currently, two approaches to lamellar replacement of the endothelium are used in the clinical setting. One approach uses the technique first described by Barraquer, 20 in which an anterior corneal flap is formed by a microkeratome to access the deep corneal stroma, and the other approach uses a limbal pocket wound to access and replace the deep stromal tissue and endothelium.
In the flap technique, Jones and Culbertson 21 used a microkeratome to create a 480-μm thick, 9.5-mm diameter hinged flap that was then retracted, and a 7.0-mm diameter trephine was used to resect the recipient stroma. A slightly larger 7.2-mm diameter donor button was then sutured into place, the flap reposited, and the flap also sutured into place. Busin et al. 22 have modified this technique and recently reported clinical results of six patients. All six patients had less than 4 diopters of astigmatism by 1 month and the authors thought that the irregularity of the astigmatism was reduced. The surface epithelium that was lost from the keratome flap took up to 14 days to reepithelialize, and they also encountered the type of flap-related complications of epithelial ingrowth and corneal melting occasionally seen with LASIK surgery. The flap approach to endothelial replacement has the distinct advantage of automated access to the mid and deep stroma and the creation of a possibly smoother interface than with manual dissection. Also, the techniques of trephination and transplantation with suturing of the donor posterior button are familiar territory to the corneal surgeon, and this approach gives easy access to other intraocular work, such as lens exchange, vitrectomy, and iridoplasty. However, the significant disadvantage of this approach is the use of sutures in the corneal tissue for the donor button and the surface flap. The compressive “donut” effect on the sutured edge of the donor–recipient interface would initially be transmitted from the interface to the surface of the flap. Also the macrostriae and microstriae induced by peripheral flap sutures are well known by LASIK surgeons, and the flap created by this procedure would not be exempt from these same compressive forces. Finally, the loss of epithelium from the surface of the flap during its relatively prolonged retraction time increases the risk of infection, ulceration, and epithelial ingrowth into the interface. Whether the advantages of this automated approach outweigh the disadvantages of suturing with this technique remains to be seen.
In 1993 Ko et al. 23 described a technique of posterior lamellar keratoplasty in rabbits through a limbal incision whereby the transplanted tissue was sutured against the recipient posterior corneal surface. Melles et al. 11 developed a technique for creating deep lamellar dissections through a limbal incision and have used this approach for deep anterior lamellar keratoplasties and most recently have shown its advantages for posterior lamellar keratoplasty. 10,12,19 The most remarkable aspect of the Melles method of endothelial replacement is that the donor tissue is able to remain in place in the posterior recipient bed without sutures. This one factor allows the entire procedure to avoid any significant suture or incisional alteration of the recipient surface corneal tissue and so preserve the preoperative corneal topography.
The study presented here has verified the inherent value of the Melles method of deep lamellar endothelial keratoplasty. Using different instruments, we have shown that a limbal wound with a deep lamellar pocket allows transfer of a donor disc with no significant impact on preoperative corneal topography and our simulated keratometry results from the Orbscan maps are similar to the keratometry readings in Melles' work. 10 Moreover, we found that the immediate stability of the transplanted tissue in this eye bank model is most likely independent of endothelial pump function and more likely the result of the inherent adhesive quality of bare stromal surfaces when pressed together and when assisted by intraocular pressure. The long-term stability of the tissue and endothelial function has also been shown in Melles' report of 1-year follow-up of seven patients. 24
There are several challenges that remain in this procedure. The manual dissection of a deep lamellar plane through a 9.0-mm scleral limbal incision remains tedious with a significant risk of inadvertent perforation. Automation of this step would allow a smoother surface and more uniform depth. The greatest technique challenge in this procedure is the intralamellar trephination and scissors excision of the recipient posterior disc supported by only an air bubble in the anterior chamber. The development of an intralamellar “punch” would allow a more symmetric and uniform recipient bed and greatly reduce the duration and technical difficulty of the entire procedure. Automation with microkeratome removal of the anterior tissue of the donor to create a smoother stromal interface of the donor disc would also improve the procedure and possibly reduce postoperative interface haze. All of these instruments and others are currently in development and should considerably reduce the difficulty and learning curve of this procedure for the corneal surgeon.
The immediate topographic advantages of deep lamellar endothelial keratoplasty over penetrating keratoplasty are considerable. However, immediate topographic success does not guarantee immediate visual improvement or stabilization. Wound healing is not addressed in this report and may adversely affect long-term topography and interface optical quality. Further clinical studies are required before large-scale adoption of this technique. If a patient's long-term topography can remain relatively unchanged from the preoperative state and the lamellar interface between the donor and recipient tissue can show long-term optical clarity, the deep lamellar endothelial keratoplasty procedure may become the new standard for the surgical treatment of endothelial dysfunction.
Mr. Chuck Hess aided in the design of these instruments and Bausch and Lomb Surgical of St. Louis, Missouri manufactured and supplied the specially designed instruments free of charge for this project. The Lions Eye Bank of Oregon and the Oregon Lions Sight and Hearing Foundation supplied the eye bank eyes used in this study.
1. Maeno A, Naor J, Lee H, et al. Three decades of corneal transplantation: indications and patient characteristics. Cornea 2000; 19:7–11.
2. Mamalis N, Anderson C, Kreisler K, et al. Changing trends in the indications for penetrating keratoplasty. Arch Ophthalmol 1992; 110:1409–11.
3. Lindquist T, McGlothan J, Rotkis W, et al. Indications for penetrating keratoplasty, 1980-88. Cornea 1991; 19:210–6.
4. Williams K, Ash J, Pararajasegaram P, et al. Long term outcome after corneal transplantation. Ophthalmology 1991; 98:651–7.
5. Riddle H, Parker D, Price F. Management of post-keratoplasty astigmatism. Cur
r Opin Ophthalmol 1998; 9:15–28.
6. VanRooij J, Christo G, Geerards A. Long term follow-up and suture related complications after penetrating keratoplasty with nylon sutures. Invest Ophthalmol Vis Sci 1998; 40:S633.
7. Confino J, Brown S. Bacterial endophthalmitis associated with exposed monofilament sutures following corneal transplantation. Am J Ophthalmol 1985; 99:111–3.
8. Wood T. Lamellar transplants in keratoconus. Am J Ophthalmol 1977; 83:543–5.
9. Sugita J, Kondo J. Deep lamellar keratoplasty with complete removal of pathological stroma for vision improvement. Br J Ophthalmol 1997; 81:184–8.
10. Melles G, Eggink F, Lander F, et al. A surgical technique for posterior lamellar keratoplasty. Cornea 1998; 17:618–26.
11. Melles G, Rietveld F, Beekhuis W, et al. A technique to visualize corneal incisions and lamellar dissection depth during surgery. Cornea 1999; 18:80–6.
12. Melles G, Lander F, vonDooren B, et al. Posterior lamellar keratoplasty for a case of pseudophakic bullous keratoplasty. Am J Ophthalmol 1999; 120:346–8.
13. Terry M, Ousley P, Zjhra M. Hydration changes in cadaver eyes prepared for practice and experimental surgery. Arch Ophthalmol 1994; 112:538–43.
14. Binder PS. The effect of suture removal on keratoplasty astigmatism. Am J Ophthalmol 1988; 105:637–45.
15. Terry M, Ousley P. Early and late astigmatism in corneal transplantation for Fuchs' dystrophy and keratoconus [abstract]. Invest Ophthalmol Vis Science 1999; 40:S632.
16. Terry M. The evolution of lamellar grafting techniques over 25 years. Cornea 2000; 19:611–6.
17. Benson W, Goosey J. Lamellar keratoplasty. In: Krachmar J, Mannis M, Holland E, eds. Cornea: surgery of the cornea and conjunctiva. St. Louis, Mosby-Year Book, 1997.
18. Arentson J. Lamellar grafting. In: Brightbill F, ed. Corneal surgery: theory, technique and tissue. St. Louis, Mosby-Year Book, Inc. 1993.
19. vonDooren B, Melles G, Lander F, et al. Preliminary clinical results of posterior lamellar keratoplasty [abstract]. Invest Ophthalmol Vis Sci 1999; 40:S632.
20. Barraquer J. The technique for penetrating keratoplasty. In: Barraquer J, Rutlan J, eds. Microsurgery of the cornea: an atlas and textbook. Barcelona: Ediciones Scribes, 1984:289–94.
21. Jones D, Culbertson W. Endothelial lamellar keratoplasty (ELK) [abstract]. Invest Ophthalmol Vis Sci 1998; 39:S76.
22. Busin M, Arffa J, Sebastiani A. EKP as one alternative for penetrating keratoplasty for the surgical treatment of bullous keratopathy: initial results. Presented at the 103rd Annual Meeting of AAO, Orlando, Florida, October 25, 1999 (program page 242).
23. Ko W, Freuh B, Shield C, et al. Experimental posterior lamellar transplantation at the rabbit cornea [abstract]. Invest Ophthalmol Vis Sci 1993; 34:1102.
24. Melles G, Lander F, vonDooren B. Preliminary clinical results of posterior lamellar keratoplasty. Presented at the 103rd Annual Meeting of AAO, Orlando, Florida, October 25, 1999 (program page 242).
25. Serdarevic O, Renard G, Pouliquen Y. Randomized clinical trial comparing astigmatism with visual rehabilitation after penetrating keratoplasty with and without intraoperative suture adjustment. Ophthalmology 1994; 101:990–9.
26. Jaffe NS, Clayman HM. The pathophysiology of corneal astigmatism after cataract extraction. Trans Am Acad Ophthalmol Otolaryngol 1975; 79:615–30.