In the past, a major problem during deep anterior lamellar keratoplasty (DALK) was the inability to obtain a smooth surface. Without it, you could not provide the patient with quality vision. This requirement may now be filled by nonmanual dissection techniques using air1,2 or viscoelastic.3,4 The combination of these adjuvants has heightened our ability to obtain a Descemet's membrane detachment with a reduced risk of intraoperative corneal perforation.5 Our DALK approach is an extension of the technique reported by Anwar and Teichmann in 2002.1
SURGICAL TECHNIQUE, TIPS
The anterior keratectomy step
We begin with a partial trephination to a depth between 50% and 75% of the central pachymetric corneal thickness (Figure 1). The partial trephination creates a circular corneal groove in which a 30.5-gauge disposable needle is inserted. Air is then injected. The needle bevel is oriented downward. Above all, we recommend maintaining a safe distance from the posterior corneal surface to limit the risk of puncturing Descemet's membrane. The purpose of this first air injection is to obtain a corneal emphysema to facilitate the anterior lamellar keratectomy. In comparison, the purpose of air injection during the original technique is to obtain a big bubble.
Nontransfixing trephination also allows us to contain the corneal emphysema, provided the needle is not inserted too deeply. Air does not diffuse within the anterior chamber through the trabeculum, and we do not systematically perform a paracentesis to control the intraocular pressure (IOP).
With a rounded blade, we then excise the anterior corneal lamella. We recommend excising no more than two-thirds of the corneal thickness, as subsequent steps are easier with a thick residual stroma—especially during the learning curve.
The nick and snake step
The originality of our technique lies in this step (Figures 2 and 3). We place a nonpenetrating stromal corneal nick slightly inside the limiting ring. The nick may be made with a sharp or rounded blade. We are comfortable using a sharp blade, which enables better control of the incision depth. We find it useful to make the nick slightly oblique and not perpendicular to the surface; its intrastromal trajectory is thus longer, and inserting the blunt cannula—by firmly grabbing the inner edge of the nick with forceps—is easier. Some pressure must be applied to initiate the path. The air emphysema facilitates cannula progression in a plane parallel to the posterior corneal surface because of the lamellar structure of the corneal stroma. We progressively advance the cannula by rotating the 10-mL air-filled syringe onto which it is mounted. The cannula appears to snake into the stromal thickness.
DETATCH THE MEMBRANE
At this point, we inject air with the cannula aperture oriented slightly downward. If the nick is deep enough, a big bubble will detach Descemet's membrane. We usually see its centrifugal progression. As Anwar and Teichmann1,2 noted, the appearance may be explosive, with a peripheral, whitish, semiopaque disc.
It is essential not to interrupt air injection until the bubble is complete. Even if the bubble progresses, air may escape along the body of the cannula or the overlying stroma. Changing the air syringe during the procedure usually results in an air leak with no further Descemet's membrane detachment. Thus, we recommend using a 10-mL airtight syringe to increase the air volume that may be injected in a single attempt. In the majority of cases, a small amount of air is sufficient. Nevertheless, when the border of the bubble does not coincide with the border of the limiting ring, viscoelastic injection obtains further detachment.
If a detachment bubble is not seen, several paths may be taken from the same nick, and air may be injected as many times as necessary. If continued air injection results in the formation of corneal emphysema without a bubble, we recommend that another manual lamellar keratectomy be performed, and another nick should be made in the residual stroma. Further attempts to achieve a Descemet's membrane detachment with the blunt cannula may then be made.
The residual stroma resection step
Once an air bubble is obtained, we exchange air for viscoelastic through the same blunt cannula (Figure 4). Whereas the blunt cannula easily progresses within the stromal lamellae, we exert some downward pressure to perforate the remaining stroma and penetrate the air bubble. The air bubble will collapse after a sudden easing of resistance.
At this point, we inject viscoelastic more slowly and in small quantities to ensure that we do not see any unusual distension of the bubble or increased IOP. If necessary—and with no disadvantage—make a sideport incision to control IOP. We do not recommend, however, making this incision before obtaining a big bubble, because it is advantageous to maintain the natural counterpressure of the eye when snaking and injecting air with the cannula.
By maintaining a safe distance with Descemet's membrane, the viscoelastic allows for a smooth removal of the remaining stroma without further dissection with a spatula. We incise the stroma along the body of the cannula, which is inserted into the bubble, and we excise it with curved microscissors along the limiting ring. It is essential to profusely wash out the viscoelastic to avoid a fluid pocket in the lamellar interface. From this point, we suture the corneal donor button—already stripped of its Descemet's membrane and endothelium—into the recipient corneal bed. It is important to take care not to perforate Descemet's membrane with the needle tip.
OBSERVATIONS
We believe that the advantages of our technique lie in (1) the reduced risk of accidental intraoperative Descemet's membrane perforation and (2) being able to attempt multiple inductions of pneumatic cleavage of Descemet's membrane during the same procedure. Because the needle tip position in the corneal thickness may be difficult to assess, we prefer maintaining a safe distance from the posterior corneal surface. Cleavage of Descemet's membrane from the posterior stroma is targeted at the subsequent nick and snake step, which may be repeated as long as the stromal thickness allows. This repeatable procedure makes it safer and reassuring, even for surgeons inexperienced with DALK. We have noticed that the number of intermediary keratectomies with subsequent nicks decreases as the surgeon's experience increases.
In the original technique,1,2 inserting the needle into the mid- rather than the deep stroma may limit the risk of puncturing Descemet's membrane, but it also limits the chance to cleave Descemet's membrane from the posterior stroma. We believe our technique can, in this case too, be used as a second choice to further increase the frequency of big bubble formation, rather than manually dissecting the corneal stroma layer by layer when a Descemet's membrane detachment is not achieved.
Finally, the use of viscoelastic ensures a complete Descemet's membrane detachment. Viscoelastic progression to the border of the limiting ring is more visible than air progression. We may completely control the limits of the bubble without further peripheral dissection with a spatula. Therefore, no further dissection is needed to reach the trephination groove or detach residual adherences between Descemet's membrane and the stroma. Additionally, we find it easier to remove the residual stroma with viscoelastic, making pre-Descemet's manipulations safer.
Because of the smooth and deep interface between the donor and recipient, and because the surgery is extraocular, postoperative outcomes are as promising as with previous techniques. Studies regarding refractive outcome and endothelial cell density are ongoing.
IN CLOSING
We believe our procedure is repeatable because of the ability to perform the nick and snake step until a big bubble is obtained, as a first or second choice, on the same or different patients. This procedure is also an accurately reproducible step-by-step technique.
Pierre Fournié, MD, is from the Ophthalmology Department, Purpan Hospital, in Toulouse, France. Dr. Fournié states that he has no financial interest in the products or companies mentioned. Dr. Fournié may be reached at pierrefournie@yahoo.com.
Jean-Louis Arné, MD, is from the Ophthalmology Department, Purpan Hospital, in Toulouse, France. Dr. Arné states that he has no financial interest in the products or companies mentioned.
