The possible advantages of transplanting only the corneal endothelium when its dysfunction is the cause of the pathology – as in corneal edema – have been considered since ancient¹. In the middle of the last century, two options for endothelial transplantation were proposed: the anterior and the posterior approach. The first description of a posterior lamellar keratoplasty is due to José I. Barraquer in 1950: after manually dissecting a rectangular anterior flap, he trephined the stroma in the bed to suture a posterior lamellar donor disc in place and finally replaced the flap². Later, he applied the microkeratome for the anterior lamellar cut (Figures 1 and 2b).

In 1956, C.W. Tillett described an analogous procedure by posterior route: dissected the deep layers of the recipient cornea and inserted a strip of stroma with endothelium from the donor in the anterior chamber (AC) to suture it in the posterior bed previously created. He used air for the first time in AC to facilitate the adhesion of the transplanted tissue³. The characteristics of the different endothelial keratoplasty techniques developed since the beginning of the century are summarized in Table 1.

ENDOTHELIAL KERATOPLASTY BY ANTERIOR APPROACH

In 1998, Jones and Culbertson, who apparently did not know about Barraquer's work, described a procedure to transplant the corneal endothelium by means of an anterior approach that they called "endothelial lamellar keratoplasty" (ELK)⁴. The donor cornea was mounted on an artificial AC, pressurized and cut with a microkeratome to separate a complete cap of half the corneal thickness and about 9.5 mm in diameter. The cornea thus prepared was trephined to 7 mm from its posterior face to obtain a disk or lenticule with endothelium, Descemet membrane (DM) and posterior stroma. In the recipient cornea, a hinge flap (LASIK type) of approximately half the corneal thickness and between 9.5 and 10 mm in diameter was carved with a microkeratome. Then a 6.75 mm window was trepanned in the posterior stroma. In this, the donor disk previously carved was sutured with independent stitches. The anterior flap was repositioned and sutured equally (Figure 2).

With this technique, some cases were published in which corneal edema was resolved, although in general only poor visual improvements were achieved. Other authors such as Busin, Ehlers, or Azar^5-7 described modifications of the same with varying success rates. It could be demonstrated, in any case, that an edematous cornea could be cleared after the transplant of a posterior corneal disc that included healthy endothelium, as well as the advantages of mechanized cutting with microkeratome.

The anterior approach was associated with various complications such as epithelial growth under the anterior flap, double AC and irregular astigmatism. Critics of this technique argued, as an additional cause for concern, the possibility (never confirmed) that these grafts could be mechanically unstable and lead to ectasia.

ENDOTHELIAL KERATOPLASTY BY POSTERIOR APPROACH

Posterior lamellar keratoplasty (PLK/DLEK)

In 1998, Gerrit Melles published the first case of restoration of corneal transparency through the replacement of its posterior layers in a patient with edematous keratopathy^8,9. Possibly because he used a posterior approach, he used this adjective in the term he proposed as "posterior lamellar keratoplasty" (PLK) (Figure 3). In the decade that followed, this type of approach became progressively preferred to treat endothelial dysfunction to the detriment of penetrating keratoplasty (PK)¹⁰.

Melles’ PLK used a scleral incision of 9 mm and a lamellar dissection with spatulas in the deep stroma, to access the DM-endothelium and trim it with scissors. This was then replaced by a donor disk obtained in the same way and with healthy endothelium. The advantages of this technique included, first, using an incision similar to what many cataract surgeons were already familiar with for the extracapsular technique. It avoided inducing a major astigmatism – one of the main disadvantages of PK – and functional recovery was faster, with fewer postoperative complications.

Surgeons from half the world made a pilgrimage to Rotterdam to learn the innovative techniques of Melles, master their “mirror”, “indentation” and “fold” signs (see chapter 5.4.2) to reach the desired corneal plane¹⁰ and familiarize ourselves with the necessary instruments. Many of us began to practice them in deep anterior lamellar keratoplasty (DALK) for corneal diseases with healthy endothelium¹¹ (see chapter 5.1). However, the technical complexity of the PLK seemed almost unattainable. Later, Melles reduced the size of the incision to 5 mm, by folding the donor disk in the form of a “Mexican taco” to introduce it in the AC of the recipient and deploy it there¹².

Shortly after, Mark Terry developed a technique in Portland, Oregon, which he called "deep lamellar endothelial keratoplasty" (DLEK), with which he treated important series of patients^13,14. The DLEK was based on Melles' PLK, but with the use of a specific instrument developed by him. Terry became, for several years, the champion of the PLK/DLEK, with a view to demonstrating his superiority with respect to PK in terms of postoperative astigmatism and early functional recovery^15-18. Although Terry's work was sometimes met with some disbelief, precisely because of his solitary personal dedication – at that time almost unique in this field – we must recognize that perhaps the PLK/DLEK did not fall into oblivion thanks to his efforts.

Descemetorhexis and microkeratome (DSEK/DSAEK/ELK)

The technical difficulty limiting the popularity of the PLK/DLEK was mainly related to the resection of the posterior corneal tissue and the creation of a donor disc congruent with said resection. Therefore, the finding of Melles in 2004, of the peeling maneuver of DM-endothelium that he called descemetorhexis¹⁹ (Figure 4), was a first fundamental milestone in the popularization of endothelial keratoplasty. The descemetorhexis greatly simplified the procedure by avoiding all the lamellar dissection maneuvers in the recipient stroma. In addition, the surface obtained by peeling was much more homogeneous than those previously created by manual dissection.

In the United States, Price's group applied the new Melles’ maneuver routinely in a technique they called Descemet's Stripping Endothelial Keratoplasty (DSEK)^20,21. However, manual dissection of the donor disc continued to be a challenge and limited the quality of the interface. The mechanization of this stage, through the use of a microkeratome on artificial AC, allowed the reproducible obtaining of donor discs of very homogeneous surfaces, which facilitated the union with the receptor^22,23.

The conjunction of both advances received a name in which the term "automated" was added to the previous DSEK, to give the today popular DSAEK (for Descemet's Stripping Automated Endothelial Keratoplasty) and explains the paradigm shift that led to the "take-off" of endothelial keratoplasty. This is, however, a terminology that primarily alludes to aspects of the technique rather than the nature of the transplant. In comparison with what would arrive later – the DM-endothelium isolates –, what defines this type of graft is the inclusion of a stromal lamina together with the DM-endothelium. For this reason, it seems more appropriate, in the current perspective, to return to the term "endothelial lamellar keratoplasty" (ELK) (see chapter 4.1).

Descemet-endothelial keratoplasty (DMEK/DEK)

Despite its success, certain aspects of the ELK (DSAEK) indicated that we were still far from perfect. The thickness of the graft could limit vision or cause problems in some patients with narrow chamber or anterior synechiae. In search of a restitution closer and closer to the anatomy, Melles would again develop the exclusive replacement of the DM-endothelium complex without any stroma. He called this technique Descemet's Membrane Endothelial Keratoplasty (DMEK)^24,25, which we also refer to as "Descemet-endothelial keratoplasty" (DEK).

Due to the minimum amount of tissue provided and maximum homogeneity of the donor and recipient, the DEK/DMEK allows to recover excellent visual acuities in periods as short as one week²⁶. It is not surprising that this new technique is progressively replacing ELK as the preferred method of endothelial transplantation. Although it is more complex from the surgical point of view, it requires less equipment and, after passing the learning phase, the results obtained are superior.

However, the ELK (DSAEK) should not disappear from the armamentarium of the corneal surgeon, due to the great experience accumulated with very satisfactory results²⁷, as well as its safety that allows applying it in situations in which the DEK has less probability of success. These include, for example, all those in which visualization of AC during surgery is limited, or others that hinder graft adhesion such as the presence of an aqueous drainage valve or other cause of hypotonia, aphakia, irregular posterior corneal surface as in PK failures, iris defects due to illness or trauma, patients with limited mobility, single eyes or cases without visual recovery capacity. On the other hand, the perspectives of endothelium culture technology and the possibility of medical therapies leave open the possibility of treating endothelial dysfunction in the coming decades.