Lamellar keratoplasty (LK), was first proposed in the first half of the 19^th century by von Walter, but it was von Hippel in 1866, that was the first to carry it out successfully¹. Although the first successful penetrating keratoplasty (PK) performed by Zirm in 1905 is often cited, most corneal transplants were lamellar until well into the first third of the 20^th century. Thus, LKs were the majority in the series with which Filatov presented in 1934, the results of his method of conservation, which allowed a viability of up to 41 hours after death².

Advances in eye banks, together with the introduction of the surgical microscope, monofilament sutures and corticosteroids as a method to prevent rejection, led to the rise of PK over the past century. The best visual results of this and its surgical simplicity relegated LK to a marginal role, which barely reached 3% of the cases despite its evident advantages, such as closed eyeball surgery, conserving the endothelium of the recipient and reducing the risk of rejection.

THE IMPORTANCE OF DEPTH

In the mid-twentieth century, some surgeons such as José I. Barraquer understood that the worst results of LK with respect to PK were related to the residual tissue in the bed. According to him, the conditions to improve visual acuity should include: achieving the deepest and most regular dissection to avoid further scarring, obtaining a graft of the highest quality and thickness possible, ensuring a good coaptation of the graft with uniformity in the traction of the sutures, and a good transparency in the interface³. Hallerman in 1959 tried for the first time to dissect up to Descemet's membrane (DM), although he placed a full thickness graft with endothelium⁴. Mc Culloch in 1963 was the first to describe the easy removal of the endothelium for full-thickness grafts over a lamellar bed⁵.

In 1965 Malbrán presented his “peeling off technique” (see chapter 5.4.1) for LK with which he obtained better results in the treatment of keratoconus and keratoglobus⁶. Other authors such as Polack, Wood or Vancea used this technique or other of manual lamellar dissection with very unequal results^7-9. Gasset addressed quite deep stromal planes and achieved final visual acuity of up to 20/30 in 80% of the cases of keratoconus¹⁰. Innovations, such as mechanized dissection with the Barraquer microkeratome, further increased the speed of the technique but LK remained below PK in terms of visual results¹¹.

Anwar first described the dissection under direct visualization of the plane of cleavage between the deep stroma and DM in 1974¹². He thus achieved a homogeneous bed, on which to place the donor button, without endothelium or DM. This improved coaptation, prevented the formation of an anterior pseudo-chamber and prevented scarring at the interface, as histological studies subsequently demonstrated¹³. However, this technique presented considerable difficulty and high risk of DM perforation, so it was not popularized.

FIRST TECHNIQUES WITH AIR AND OTHER FLUIDS

The first technique that we can consider within the current concept of deep anterior lamellar keratoplasty (DALK) is due to Enrique Arenas Archila, who proposes, in 1985 in Colombia, the injection of 1 ml of air through a 26 G needle, just above of the DM. He then performs trephination and dissection with spatula of the stroma until DM – which will be the only layer not infiltrated with air, to graft a full-thickness cornea with endothelium, and he apparently did not find that causing problems¹⁴. He also recommended performing a paracentesis to facilitate dissection in hypotonia, but not before the injection of air as this could pass to the anterior chamber through the puncture. With this technique, Price found in 1989 the need to convert to PK in 3 of 10 cases, for perforation of the DM¹⁵.

Sugita and Kondo published in 1994 a series of 120 eyes with a DALK technique by "hydro-delamination": after trepanning to 2/3 of depth and dissecting the stroma with Desmarres’ knife and Pauique’s spatula, they injected saline into the bed with a 27 G needle and resected the residual stroma with a 0.25 mm spatula. Although they had 39% of DM perforations, they managed to seal them with air in the anterior chamber. In no case was there an immune rejection of the graft¹⁶. Morris et al modified this technique by adding viscoelastic after the saline injection, as well as a paracentesis that would prevent the rupture of the DM during the resection of the last layers of the stroma. The subsequent study with specular microscopy did not show a significant loss of endothelial cells¹⁷. Amayen and Anwar also test hydro-delamination, but saline does not achieve proper diffusion in areas with deep scarring and only reaches the level of DM in 11.5% of cases¹⁸.

In 1998, Tsubota et al describe a technique called "divide and conquer", inspired by the namesake for cataract surgery. After trepanation of 70% and lamellar dissection, the stroma is divided into 4 sectors, which facilitates resection. They apply a simple self-adjusting continuous suture, during and after surgery, to minimize residual astigmatism. No DM perforation was reported, and the mean residual astigmatism was 2.5 diopters at 6 months. However, the final visual acuity levels were low, which is attributed to irregularities in the interface because it is a manual dissection¹⁹.

Among the advances that have most stimulated the rebirth and current popularity of the lamellar techniques, several contributions by Gerrit Melles stand out. In the field of DALK, he devised a method that uses air, not as an intrastromal dissector, but by filling the anterior chamber to visualize the depth of the dissection. Using spatulas of his design inserted from the limbus, its reflection in the surface of the bubble allows calibrating the depth, and thus creating a predescemetic pocket, without the need for a layer-by-layer dissection. With this technique, he achieves results comparable to PK in terms of visual acuity and astigmatism²⁰. He also proposes the injection of viscoelastic in the pocket to separate adhesions of the stroma and avoid perforation during trepanation, a complication that can be reduced to 12%²¹.

THE TECHNIQUE OF THE "BIG BUBBLE" AND ITS VARIANTS

In 2002, Anwar and Teichmann described and named what has become the reference technique for DALK²². Perhaps the resonance or alliteration of the term "Big Bubble" (BB) contributed to its success, or it was just the spectacular sudden effect of pneumo-dissection that can be obtained. But it is undoubtedly the most elegant and quick procedure to reach the plane before the DM – as long as the BB is formed. Unlike Arenas’ technique, partial trepanation is done before the air injection. The BB technique is described in detail in chapter 5.4.3.

The difficulty to obtain a BB in all cases has been a source of variations and improvements. Donald Tan's group introduced the use of a small bubble in the anterior chamber as an indicator of the success of the formation of the BB, since the latter creates a central depression of the DM, which will displace the small bubble to the periphery of the anterior chamber²³. By removing a first sheet of medium thickness, prior to the injection of air through an incision in the stromal bed, Fournie et al reduce the perforations to 15.3%²⁴, while J. Behrooz and Daneshgar²⁵ propose, as "large bubble", to trepan only 20-30% of the thickness and to inject the bubble at a superficial level²⁵. In contrast, the "small bubble" of Busin’s group consists of making a limited injection of air, which separates only the central 5-6 mm of the DM, while in the periphery a layer of deep stroma is left which is somewhat thicker, which would avoid perforations²⁶.

Another critical aspect is the injection of air, and deciding its most effective level, to create the BB. This has led to several special instruments (spatulas, cannulas), such as Sarnicola’s, with which BB is successful in 26 of 28 cases²⁷. Scorcia et al take as a reference the reflection of the Fontana’s cannula on the background glare²⁸, while the group of Mc Kee proposes injection of triamcinolone in the anterior chamber²⁹. The use of intraoperative optical coherence tomography to identify the planes has even been described³⁰.

THE FUTURE: TECHNOLOGY OR ART

The development of refractive surgery in the second half of the 20^th century – especially in the last quarter – is among the factors that have most influenced the restoration of LK and its current boom. Through it the corneal surgeon has become familiar with the creation of beds and manipulation of sheets of corneal tissue (PRK, LASIK, SMILE, etc.). It is not surprising that technological means linked to refractive surgery, such as microkeratomes and different lasers, that have also been applied to the field of LK.

Microkeratomes have become a routine instrument in superficial anterior lamellar keratoplasty (SALK) and in obtaining grafts for endothelial lamellar keratoplasty (DSAEK). Positive experiences have been reported with its use for LK of medium thickness (donor disk cut at 350 μm and recipient bed at 250 μm) in cases of keratoconus³¹, as well as with excimer³² or femtosecond laser³³. However, as discussed in Section 4, these technologies have proved their efficacy and safety in a contrasted way only in the cases of the most superficial pathologies, in which they do not surpass 200 μm of depth. The results have been at least variable in the face of deeper injuries, and it is not clear that they surpass those of the manual techniques of the classic LK or those of the current DALK.

The historical evolution of DALK is an example of the triumph of essentially manual techniques. Attempts to apply the femtosecond laser – beyond the preliminary phases of the intervention, such as obtaining the graft, partial trepanation of the recipient or creation of an initial plane of medium dissection – have not consistently achieved the level of dissection close to the DM that is required for a DALK, and it has been argued that laser energy would cause greater inflammation and a greater risk of rejection³⁴. Only time will tell us if one day technology will overcome the art of pneumo-dissection.

At present, the anterior LK (in its variants DALK and SALK) has established itself as a solid alternative to PK – and for many surgeons the first, in practically all of the corneal pathologies that do not affect the endothelium. Some controversy persists regarding the visual results: while the American Academy of Ophthalmology has concluded that they are equivalent in both techniques, the Australian Corneal Graft Registry disagrees. This may depend on aspects that vary among surgeons or cases. However, before a possible tie, the obvious structural advantages of a less invasive, more selective and less risky surgery tips the balance. Only the difficulty (either manual or mental) in the learning of the techniques explains that – unlike what happened with endothelial keratoplasty – anterior LKs have not yet displaced PK to a greater extent.