In the past several years, lamellar keratoplasty methods have been developed as alternatives to penetrating keratoplasty (PK) for the treatment of corneal disorders.
In corneal states such as Fuchs' endothelial dystrophy and aphakic and pseudophakic corneal edema, the pathologic decompensation of the endothelial layer causes epithelial and stromal edema and subsequent severe visual loss. PK—the procedure of choice for the treatment of the symptomatic stage of these conditions—is often complicated by high or irregular astigmatism, suture-related problems of ulceration and vascularization, graft rejection, and insufficient wound healing, which may result in wound dehiscence or rupture after minor ocular trauma.
Technological and surgical advancements, as well as a greater understanding of corneal physiology and optics, however, have led to more progressed and refined endothelial keratoplasty procedures. These lamellar surgical techniques have achieved visual results approaching those of PK, and they have reduced the rejection rates and improved long-term graft stability.
LAMELLAR REPLACEMENTS
Currently, there are two clinical approaches to lamellar endothelium replacement. The first approach uses a technique first described by José I. Barraquer, MD, of Spain, in which an anterior corneal flap is formed with a microkeratome to access the deep corneal stroma. In the second approach, a limbal pocket wound is used to access and replace the deep stromal tissue and endothelium. The concept of the posterior approach to endothelial keratoplasty was first described by Ko et al1 in 1993. In 1998, Melles et al2 described a method for deep lamellar dissection through a limbal approach that was previously used for deep anterior lamellar keratoplasty.3 In this technique, a 9-mm scleral incision is used to create a stromal pocket across the cornea at approximately 80% stromal depth. A 7.5-mm posterior lamellar disc (consisting of the posterior stroma, Descemet's membrane, and endothelium) is excised and replaced with donor tissue of the same size.
Melles et al2 reduced the length of the scleral access incision from 9 to 5 mm in an attempt to accelerate visual rehabilitation and provide better wound strength and trauma resistance.4 They also presented laboratory results from three human cases in which Descemet's membrane was stripped from the recipient to provide a smoother recipient interface. This step provided a bed for the donor tissue to stick directly onto the posterior surface and eliminated stromal dissection from the deep lamellar endothelial keratoplasty (DLEK) procedure.2,4 Francis W. Price Jr, MD, of Indianapolis, was the first surgeon to publish clinical results for this type of descemetorhexis technique. Later, he renamed the method Descemet's stripping endothelial keratoplasty (DSEK).5 In both DLEK and DSEK, the recipient corneal topography is preserved, and little astigmatic change is induced postoperatively.5,6
CORNEAL TISSUE TRANSPLANTATION
Jones and Culberston7 introduced posterior corneal tissue transplantation using the anterior approach. They used a microkeratome to create a 480-µm–thick, 9.5-mm hinged flap. The flap was then retracted, and a 7-mm trephine was used to resect the recipient stroma. A slightly larger 7.2-mm donor button was sutured into place, and the flap was reposited and sutured. Busin et al8 modified this technique and reported clinical results for six eyes. In their technique, the flap thickness was reduced to 160 µm, with a 6.5-mm central posterior lamella in the recipient and a 7-mm donor button. The donor button was sutured into position using four cardinal sutures and a running eight-bite antitorque suture. The flap was then closed using a running suture.
OUR METHOD
In our study,9 we modified the surgical technique by positioning the donor tissue in the bed without sutures. In eight eyes, we used an automated microkeratome to create a 130-µm flap; we used a 250-µm flap in two eyes. The ALTK system and a LSK-One microkeratome (both manufactured by Moria, Antony, France) were used to create a centered, nasally hinged 9.5-mm flap. We retracted the flap and excised the central posterior button using a 7-mm trephine and corneal scissors.
To prepare the donor button, the donor corneoscleral rim was placed in an artificial anterior chamber (Moria), and the anterior lamella—as thick as the flap—was removed with the microkeratome. The remaining donor tissue was inverted so that the endothelium surface was turned upward on the trephine block to obtain a 7-mm button. The anterior chamber was filled with viscoelastic, and the donor lenticule was placed in the recipient bed; the flap was replaced and properly aligned. It was then sutured in position using eight or nine interrupted 10-0 nylon sutures.
At the end of the procedure, the viscoelastic was removed from the anterior chamber through a paracentesis incision and was replaced with two parts air and one part balanced salt solution. The main difference in our modified technique was the sutureless fixation of the posterior donor button in the recipient bed. We only found one other report of a sutureless technique.10 Melles2 first introduced the concept of using intraocular air to support the graft tissue and sutureless donor button fixation in a DLEK technique. He named this procedure posterior lamellar keratoplasty (PLK)—a name that is now used by many surgeons who perform the flap technique of endothelial replacement surgery. Unlike previous studies, our technique used a donor button identical to the 7-mm trephination in the host stromal bed. In a laboratory model for microkeratome-assisted PLK,11 it was found that the larger the button size was in relation to the trephined recipient bed, the more astigmatism was induced on the corneal surface. The same size button, however, resulted in a greater hyperopic shift by inducing corneal flattening. The larger-sized buttons had less effect on the keratometry reading.
In our study, all uncomplicated cases had stable refraction and corneal topography as early as 1 month after surgery, which was maintained during follow-up. In these eyes, all sutures could be safely removed 3 to 5 months postoperatively. Busin et al8 found no major refractive changes after suture removal. In all patients (except one patient who had undergone endothelial regraft surgery), regular astigmatism was evident on corneal topography and keratometry. Corneal astigmatism ranged from 2.25 to 6.50 D; it exceeded 5.00 D in two of eight eyes. One month postoperatively, the lowest UCVA in pseudophakic eyes was 20/400, and the BCVAs ranged from 20/400 and 20/25. These results could be attributed to the accompanying complications of a previous cataract surgery, pseudophakic and aphakic corneal edema, the state of the macula, and glaucomatous damage of the optic nerve.
IMPORTANT ADVANCES
The technological advancements in instrumentation have been critical in the evolution of the anterior approach technique. The popularization of refractive surgery and major improvements in microkeratome technology have allowed for more precise donor and recipient resection and smoother interface compared with manual dissections. This may help minimize interface irregularities in the early postoperative period as well as produce surfaces of excellent optical quality. It seems that in the wide central area dissected by the microkeratome, postoperative healing time and scar formation are minimized tangentially to the corneal surface. In the peripheral annular area of contact between donor and recipient stroma, however, scarring similar to that observed during normal corneal wound repair occurs. As a result, donor lamellae are safely secured to the underlying recipient cornea by the peripheral scar, and like post-LASIK outcomes, the interface plane remains optically clear with a high probability of 20/20 vision.
Another important advancement is the preparation of the donor button with the artificial anterior chamber. When compared with the entire eye, this preserves the endothelium of the corneoscleral cap and gives greater flexibility for lamellar endothelial replacement. Various modifications and improvements in the artificial anterior chamber help ophthalmic surgeons to trephine to any desired donor corneal depth. The artificial anterior chambers may be used for both manual or automated lamellar dissections of the cornea using a microkeratome and the Moria ALTK system. In the ALTK system, the high-speed, high-power turbine (30,000 cuts/minute) creates a smooth keratectomy for a seamless-edge margin. We used the Moria ALTK system, which utilizes a different design for its artificial anterior chamber, to resect the donor tissue in the following steps. First, the donor corneal cap is sealed within the artificial anterior chamber, and the intrachamber pressure is set at the required level. Second, the surgeon selects the desired diameter of the cut, and finally, the donor corneal resection is performed with the LSK-One microkeratome. The single-piece construction of the microkeratome heads are precalibrated for various depths of cut (130–400 µm).12
One of the main advantages of microkeratome-assisted PLK has over hand-dissected DLEK is its standardization. Precision and repeatability of these devices, however, have not been fully optimized, and a certain degree of variability is still present, especially in the thickness of the lamellae obtained. The current dissection depth error associated with a microkeratome is approximately 70 µm.13 Additionally, different degrees of tissue hydration at the time of dissection of both the donor and recipient corneas may strongly affect the final thickness of the dissected lamellar buttons. To avoid the negative effect of variations in the lamellar disc width on the surgery, we suggest preparing the recipient bed first and then adjusting the dimensions of the donor lenticule to the required diameter.
FLAP THICKNESS AND OUTCOMES
The optimal thickness of the retained anterior flap must be determined. A thin flap would probably clear more quickly, whereas a thicker flap could offer more stability to the corneal surface. In our study, the flap thickness was 130 µm in eight eyes and 250 µm in two eyes. It seems that a thicker flap is desirable, because a thin flap may be associated with more technical difficulty in suturing and complications (eg, flap wrinkling). We found no clinical difference between the flap thicknesses in the two groups.
Compared with a full thickness keratoplasty, PLK may have several advantages including less surgical time, a lower risk of intraoperative complications and high astigmatism, faster visual recovery, fewer follow-up visits for selective suture removal, and less side effects of topical steroid therapy. For experienced LASIK surgeons, microkeratome-assisted PLK may be relatively quick and easy to perform. Also, trephination and transplantation with the flap sutured to the recipient tissue is a familiar method for any corneal surgeon. This approach gives access to other intraocular procedures (eg, lens exchange, vitrectomy, iridoplasty). Additionally, endothelial injury, which is possible during the donor disc positioning in DLEK, is avoided in the anterior approach technique. The significant disadvantage of this approach, however, is the use of sutures in the corneal tissue for the surface flap. Also, LASIK surgeons are well aware of the macro- and microstriae induced by peripheral flap sutures, and the flap created by this procedure is not exempt from these same compressive forces. Finally, the loss of epithelium from the flap during its relatively prolonged retraction time increases the risk of epithelial ingrowth into the interface. We saw this occur in two eyes. Other reports of interface epithelial ingrowth also exist in the literature.8,14
A drawback of endokeratoplasty is the limited amount of transplantable endothelium. The transplanted posterior graft, as small as 7 mm, exchanges only 40% of the diseased endothelium. Also, corneas with long-standing edema causing superficial fibrotic changes cannot be treated with the posterior lamellar graft procedures.
Restoration of the normal corneal topography is an inherent advantage of DLEK, because surface corneal incisions and sutures are avoided.15 If a patient's long-term topography can remain relatively unchanged postoperatively and the lamellar interface between the donor and recipient tissue can show long-term optical clarity, DLEK may be the ideal surgical treatment for endothelial dysfunction. Currently, this technique is labor-intensive with a difficult learning curve for recipient disc excision. Also, the transfer of the donor endothelium in DLEK requires a different technique and skill set than traditional PK.
Furthermore, questions persist about the short- and long-term viability of the donor endothelial cells. In a prospective study that evaluated endothelial cell loss 2 years after DLEK, Terry et al16 found an acceleration of cell loss from 1 to 2 years postoperatively. This was after minimal change from 6 to 12 months postoperatively. The small-incision DLEK technique, which involves folding the donor tissue, results in a higher endothelial cell loss at 1 and 2 years compared with large-incision DLEK surgery with unfolded tissue.
In the first 2 years, the corneal donor endothelial preservation after DLEK, however, seemed to be as good as or better than preservation after standard full-thickness PK.16
OUTCOMES FOR DSEK
Although recently introduced, DSEK is evolving. DSEK consists of stripping Descemet's membrane and endothelium from a recipient cornea and transplanting the posterior stroma and endothelium of a donor cornea through a 5-mm incision. Early outcomes in the study by Price and Price5 demonstrated rapid visual recovery and minimal changes in refractive error. This technique maintains the structural integrity of the cornea by preserving the recipient's entire corneal thickness, except the endothelium. As described in recent studies, DSEK is technically easier to perform than other posterior graft techniques and the refractive outcomes are more predictable.
The donor adherence, however, is more challenging. Dislocations seem to be more frequent after DSEK compared with PLK and DLEK. This may be due to the exquisitely smooth exposed corneal surface present after Descemet's membrane stripping. Compared with the rougher stromal surface created by lamellar dissections, this surface provides less traction for donor attachment.
FEMTOSECOND TO PREPARE THE CORNEA
Some technical innovations may further improve the visual outcome of endothelial keratoplasty procedures. In recent years, the femtosecond laser has been used to cut a corneal flap at the depth of 100 to 160 µm. Excellent visual results without optical interface problems have been noted after creation of a superficial lamellar interface.17,18 It seems that the femtosecond laser can create corneal flaps in PLK and prepare donor buttons in all endothelial keratoplasty procedures in automated and standardized fashions. Endothelial cell loss after preparation of 150- to 200-µm posterior stromal discs with the femtosecond laser was approximately 4% and did not appear to be caused by laser pulse energy.19 Unlike a microkeratome posterior corneal disc, which is prepared at the time of transplantation, a button prepared with the femtosecond laser may be cut up to 3 weeks beforehand.20 This advance preparation prevents intraoperative situations in which an intended endothelial keratoplasty procedure has to be converted to PK because of complications that may occur during manual or microkeratome preparation of the donor disc. The promising feature of femtosecond laser technology in corneal endothelial transplant surgeries has been shown in a case report of femtosecond DSEK.21 Another innovation in this field is the use of adhesives, instead of sutures, to secure the flap in place.22
The most prominent advantage of endothelial keratoplasty versus conventional corneal transplantation is better and more stable refractive outcomes. Surgeons also avoid significant refractive shifts found with conventional surgery, as the full-thickness corneal wound heals and sutures are removed over the course of several years. The possibility of lifting the flap to correct residual refractive error using an excimer laser is a specific theoretical advantage for microkeratome-assisted PLK surgery. Additionally, donor tissue may also be used more efficiently. Otherwise discarded corneas, for example, may be used after photorefractive keratectomy for endothelial graft procedures. Also, the deep location of the donor button might lessen the risk of immunization of the host against the graft, and therefore lower the chance of an immune reaction. Ease of endokeratoplasty procedures may be further enhanced by the widespread use of donor tissue that has been precut by the distributing eye bank before delivery to the surgeon.
IMPROVED APPLICATION, RESULTS
Instrumentation (eg, artificial anterior chambers, advanced microkeratomes, excimer lasers) has improved the application and results of this surgical technique. As additional follow-up is obtained, complications, endothelial viability, and long-term visual results will be determined. Once the procedure is made fully practical and uniform, a large, prospective, randomized clinical trial comparing microkeratome-assisted PLK with conventional PK and other forms of endothelial graft procedures will be needed to fully establish the best standard of care for endothelial replacement surgery.
Hassan Hashemi, MD, is Head of Noor Ophthalmology Research Center, Noor Eye Hospital, and the Farabi Eye Hospital, Tehran University of Medical Sciences, in Tehran, Iran. He states that he has no financial interest in the companies or products mentioned. Dr. Hashemi may be reached at tel: +98 21 88651515; fax: +98 21 88651514; or hhashemi@noorvision.com.
Jila Noori, MD, is a researcher at the Noor Ophthalmology Research Center, Noor Eye Hospital, in Tehran, Iran. She states that she has no financial interest in the companies or products mentioned. Dr. Noori may be reached at jnoori@razi.tums.ac.ir.
