The surgical approach to the treatment of corneal endothelial disease such as Fuchs endothelial dystrophy has undergone ground-breaking changes. For more than 100 years, patients with corneal edema as a result of endothelial keratopathy were offered penetrating keratoplasty (PKP) as the only surgical treatment for their condition.¹ This approach did indeed replace the diseased cornea, but it was fraught with many problems. Penetrating keratoplasty resulted in an optically clear cornea with excellent corrected visual acuity although often after a delayed period of visual recovery, requiring further intervention (i.e., astigmatic keratotomy or selective suture lysis) and with an unstable and unpredictable refractive outcome.^2,3,4 Furthermore, the final visual outcome was best corrected visual acuity with most patients requiring correction of high corneal astigmatic changes.^2,3 Worse yet was the unpredictability and instability of the refractive outcome with significant topographic and refractive changes when sutures broke or were lysed even years later.⁴

The same full-thickness PKP corneal wound that led to the refractive limitations mentioned also limited and weakened the tactile strength of the cornea. A post-PKP cornea had an increased lifetime risk of wound rupture and dehiscence with such complication reported to occur many years after initial surgery.^5,6,7 Additionally, the sutures needed for the full-thickness grafts could be a nidus for infection or inflammatory reaction leading to graft rejection.^7,8,9

It was in 1956 when Tillet¹⁰ first described a customized approach to the surgical treatment of corneal endothelial disease and thus posterior lamellar keratoplasty (PLK) first entered the clinical domain. Tillet’s approach was through an extensive anterior lamellar dissection, trephination of the posterior cornea, transplantation of donor posterior cornea, finished with interrupted anterior corneal sutures. Although this was a more customized surgical approach to endothelial keratopathy, it did not solve many of the drawbacks of PKP inherent in needing surface sutures. It was not until 1993 when Ko et al¹¹ described at ARVO a new technique of endothelial keratoplasty done through a scleral limbal incision circumventing the need for any surface corneal sutures. Through their animal studies, they rekindled the interest in the selective approach to endothelial keratoplasty, but their technique still required sutures to attach the endothelial donor graft in place. Then Gerrit Melles et al^12,13 of the Netherlands introduced the concept of using an air bubble in the anterior chamber to hold the donor graft in place, avoiding any surface corneal incisions or sutures. In 1998 they reported on the results of the first human patient on whom Melles performed this form of selective endothelial keratoplasty and named it posterior lamellar keratoplasty (PLK).¹² In 1999 Terry and Ousley¹⁴ began laboratory work on endothelial keratoplasty, modified the technique and instrumentation, and proved the safety and utility of Healon during the initial steps of the procedure, all in hopes of making the procedure safer and easier. They then performed the first U.S. case of endothelial keratoplasty in 2000 and renamed the procedure deep lamellar endothelial keratoplasty (DLEK) to differentiate it from the original flap technique of EK, which was called PLK.¹⁵ Terry then went on to investigate EK in the largest institutional review board (IRB)-approved prospective study in the world and has through numerous published reports proven its advantage over traditional PKP as a surgical treatment for endothelial keratopathy. He has proven that with the elimination of surface corneal sutures and incisions, more predictable corneal topography, faster wound healing, more rapid visual rehabilitation, and more stable refractive outcomes can be obtained.^{15,16,17,18,19,20,21,22,23}

In 2003, further modifications were made in the technique when Melles et al²⁴ introduced descemetorhexis as a way of preparing the recipient cornea, avoiding the cumbersome step of manual lamellar dissection of the posterior corneal stroma needed in preparing the recipient cornea in DLEK. The stripping of the Descemet membrane as compared to the manual intrastromal lamellar dissection of the posterior cornea also yielded a smoother interface and likely a more optically clear cornea. Frank and Marianne Price²⁵ popularized this technique by publishing retrospective results of the first large series in 2005 and 2006 and named the procedure Descemet’s stripping endothelial keratoplasty (DSEK). To further simplify the technique, Mark Gorovoy used an automated microkeratome for preparing the donor cornea and named this new modified technique Descemet’s stripping automated endothelial keratoplasty (DSAEK), which is now the most commonly performed form of EK.²⁶

The indication for EK is corneal endothelial disease. Fuchs dystrophy and pseudophakic bullous keratopathy constitute the great majority of such cases that are commonly considered for EK surgery.²⁷ Other rare causes of endothelial disease such as iridocorneal endothelial (ICE) syndrome or traumatic endothelial cell loss constitute a much smaller fraction of cases undergoing EK.

In the preoperative evaluation of patients with corneal endothelial dystrophy, attention needs to be paid to several factors to help with the surgical planning as well as proper patient counseling on expected outcomes. The status of the lens is an important factor to consider. If the patient is pseudophakic, the stability of the intraocular lens (IOL), location of the IOL, and posterior chamber (PCIOL) or anterior chamber (ACIOL) need to be assessed. If there is an ACIOL, the decision needs to be made whether the lens can safely be left in the eye or if it needs to be exchanged for a PCIOL, either placed in a capsule with sufficient support or fixated to the sclera. A newer-model multiflex open-looped ACIOL that is stable and has been in place for years prior to the decompensation of the patient’s corneal endothelium can potentially be safely left in place without consequence.²⁸ This decision is made by the surgeon and can be a safer alternative to explanting and exchanging the ACIOL for a scleral-fixated IOL. An aphakic eye with little visual potential but with painful bullous keratopathy presents another challenge. If, in fact, the visual potential is limited at best, placing a secondary IOL would present greater risk than any benefit, and thus such unnecessary extra procedure should be avoided. On the other hand, such an eye could benefit from EK for the treatment of the painful bullous keratopathy, and the choice of proceeding with EK versus PKP may be made to avoid suture-related complications of PKP. When an ACIOL is to be left in the eye or especially when an eye is to be left aphakic, we recommend performing deep lamellar endothelial keratoplasty (DLEK) instead of DSAEK as an anterior chamber bubble cannot be maintained in these eyes and thus an increased risk of graft dislocation with DSAEK is likely. Even more devastating than graft dislocation in an aphakic eye is the donor graft “falling” into the posterior chamber. DLEK grafts do not require a persistent anterior chamber bubble to support the donor graft and have a lower chance of dislocating.²⁹ This recommendation to choose DLEK as the procedure of choice applies to all eyes in which an anterior chamber air bubble cannot be maintained such as in an eye with a large iridotomy and/or iris defects. It is beyond the scope of this chapter to describe the steps in performing DLEK, and we recommend referring to detailed published articles for an in-depth discussion and description of technique.²¹

Another point to consider preoperatively is the degree of anterior stromal opacity or scarring as a result of long-term endothelial keratopathy; this may be evident on slitlamp biomicroscopy or may also be subclinical.³⁰ A cornea that has suffered from significant and long-term bullous keratopathy, for example, will be detumesced from EK but will likely not gain the same amount of visual recovery as a cornea with little anterior stromal haze/scarring. This finding should not preclude the surgeon from offering EK to the patient and helping the patient understand the pros and cons of EK versus PKP in his or her specific case. In such cases PKP may offer a better final corrected visual acuity but again at the cost of delayed visual recovery, possible suture-related complications, and significant corneal astigmatism.

Other preoperative findings that need to be noted and assessed are the presence of prolapsed vitreous, macular or retinal pathology, pre-existing glaucoma, and other ocular pathologies that either need to be addressed prior to proceeding with EK or considered in preparation for EK and the postoperative management.

A surgical video of our DSAEK technique using “precut” donor tissue will accompany this chapter. In the paragraphs that follow, we summarize the technique.

The procedure is typically performed with a retrobulbar or a peribulbar anesthetic injection. The pupil is dilated preoperatively if concurrent cataract extraction is to be done. A 5.0-mm-long incision at a 350-micron depth (guided by a trifaceted, guarded diamond knife) is created temporally and approximately 1.0 mm posterior to the limbus and is tunneled toward and into 1.0 mm of clear cornea. Paracentesis ports are created on either sides of the scleral-limbal tunneled incision, and the aqueous is exchanged with Healon (Advanced Medical Optics, Santa Ana, CA), a cohesive viscoelastic, to stabilize the anterior chamber.

If cataract extraction is to be performed, it can be done through this scleral-limbal tunnel by creating an entrance wound of the surgeon’s desired length and proceeding with cataract extraction as usual. Dispersive viscoelastic agents such as Viscoat (Alcon, TX, USA) should be avoided during any and all steps of the procedure whether or not concurrent cataract extraction will be performed as this type of viscoelastic can adhere to the endothelium or the stroma and interfere with donor graft adherence. Another caveat is avoiding creating too large a capsulorrhexis as an unstable and poorly seated intraocular lens can be dislocated during the step of graft insertion. After lens implantation, the cohesive viscoelastic is not irrigated out of the anterior chamber, and the surgeon proceeds with the steps required to perform the endothelial keratoplasty.

A template circular mark of desired diameter (typically 8.0–9.0 mm) is placed centrally on the corneal epithelial surface, keeping in mind to clear the entrance of the scleral-limbal incision and the paracentesis ports. This mark is accentuated using gentian violet ink marks and used as a reference for the next step. With the chamber stabilized using the cohesive viscoelastic, a blunt-tip instrument such as the Terry-Sinskey hook (Bausch and Lomb, St. Louis, MO) is inserted into the anterior chamber through one of the two paracentesis ports and pointed toward the endothelium to score a circular periphery of the diseased and thickened endothelial cell layer and the Descemet membrane. The same blunt-tip hook is then used to peel the Descemet membrane gently away from the underlying stroma and pull the edges of the circular scored tissue centrally. This tissue can then be removed using Utrata forceps or the irrigation/aspiration handpiece. To lessen the higher rate of graft dislocations seen in DSAEK owing to the smoother recipient bed stripped of its Descemet membrane, we propose roughening of the peripheral 1.5 mm of the recipient bed using a specialized instrument, the Terry scraper (Bausch and Lomb, MO, USA), to expose stromal fibrils that possibly encourage better donor adhesion. After the peripheral scraping, the scleral-limbal wound is then extended to its full 5.0 mm and bisected using an interrupted 10.0 nylon or Vicryl suture. The irrigation/aspiration handpiece is entered through the wound, and thorough aspiration of the viscoelastic is done. Pupillary miosis is obtained using Miochol-E (Novartis, North Ryde, Australia), and then the surgeon’s attention is turned to preparing the donor tissue.