This article is from January 2007 and may contain outdated material.
Descemet’s stripping automated endothelial keratoplasty (DSAEK) is an evolving technique that is rapidly gaining in popularity. Advanced disease of the corneal endothelium has traditionally been treated with penetrating keratoplasty (PK), but new surgical procedures such as DSAEK may provide better patient outcomes.
Because DSAEK introduces significantly less postoperative astigmatism and has a potentially shorter recovery time compared with traditional PK, the new procedure is likely to become the standard of care for diseases of the corneal endothelium.
Corneal Pathophysiology
The average thickness of the normal adult human cornea is approximately 540 µm, and the cornea consists of five layers from anterior to posterior: epithelium, Bowman’s layer, stroma, Descemet’s membrane and endothelium. The composition of the stroma is not uniform; the anterior stroma contains a higher ratio of dermatan sulfate to keratan sulfate, making the posterior stroma more likely to swell with excess water in states of endothelial dysfunction.1
Corneal deturgescence is maintained by endothelial cell sodium/potassium-activated adenosine triphosphatase (Na+, K+-ATPase) and by tight junctions between the endothelial cells that limit the ingress of fluid. By removing fluid from the stroma and limiting its entry, endothelial cells maintain the ordered arrangement of collagen and preserve the cornea’s transparency. In states of deficient endothelial cell density, a lack of tight junctions between the endothelial cells allows increased entry of fluid into the stroma. The endothelial cells that remain may have a higher concentration of Na+, K+-ATPase as a compensatory mechanism to increase fluid removal.1
The cornea generally has an endothelial cell density of 5,000 to 6,000/mm2 at birth, a number that decreases to approximately 2,500 to 3,000/mm2 by early adulthood. Then, after that, there is an average cell loss of 0.6 percent per year.1 Edema develops when endothelial cell density drops below 400 to 700/mm.2
Because the endothelial cell density is higher in the periphery of the cornea, it is thought that the peripheral cells may serve as a reserve for those lost centrally.1 Human corneal endothelial cells are arrested in the G1 phase of the cell cycle; it is believed that they undergo few, if any, mitoses in vivo.2 Serious derangement of the endothelial cell layer, therefore, results in irreparable pathology that requires corneal transplantation.
In traditional PK, all layers of the cornea are replaced. This is unnecessary in diseases like Fuchs’ dystrophy and pseudophakic bullous keratopathy, which are disorders of the endothelium. DSAEK removes only the endothelium and Descemet’s membrane, replacing these layers with donor endothelium, Descemet’s membrane and posterior stroma.
Performing DSAEK
DSAEK is evolving rapidly, and many variations of the procedure are currently in use. In this article we present a generalized approach.
Patient selection. Surgeons preparing to complete their first few DSAEK procedures should select patients with deep anterior chambers. Pseudophakic patients with deep anterior chambers and posterior chamber intraocular lenses (PCIOLs) make good surgical candidates, because there is adequate space to unfold the donor button and no risk of trauma to the crystalline lens. In a patient with an anterior chamber IOL, exchanging the IOL for a PCIOL prior to DSAEK will make the procedure more manageable.
Many patients will require both corneal transplantation and cataract removal. It currently appears advantageous to perform the cataract procedure before the DSAEK, in the process creating a deeper anterior chamber and avoiding the risk of damaging the endothelium of the donor graft. Some authors have suggested a “triple procedure” involving cataract removal, placement of an IOL and endothelial keratoplasty. A potential drawback of the triple procedure is that the presence of viscoelastic on the donor stromal bed could interfere with graft adherence, but this risk may only apply to dispersive viscoelastic materials.
Graft preparation. We use precut tissue prepared as follows: The donor corneoscleral rim is positioned on an artificial anterior chamber, endothelial side down, and pressurized to approximately 90 mmHg. A lower pressure may be used for corneas thinner than 550 µm. The corneal thickness is measured carefully using ultrasound pachymetry while the artificial anterior chamber is pressurized.
The anterior corneal lamellae is removed using an automated microkeratome. We use a 350 µm head for corneas thicker than 570 µm and a 300 µm head for thinner corneas. The thickness of the residual bed, containing stroma, Descemet’s membrane and endothelium, is then measured with ultrasound; a graft thickness of 100 to 120 µm provides enough rigidity for easy manipulation, yet is thin enough for insertion into the anterior chamber through a small incision. A small dot made with gentian violet may be placed centrally on the stromal side of the graft to aid with positioning, and a button is typically cut with a trephine between 7 and 9 mm in diameter. The edges of the button are also marked to aid in graft positioning. To lower the incidence of postoperative corneal edema and graft dislocation, it is important that the donor button be sized to closely fit the area stripped of Descemet’s membrane in the recipient eye.
Note that after the anterior stroma has been removed from the donor cornea, it is difficult to distinguish the endothelial and stromal surfaces of the donor button. Gently inscribing a small “S” on the stromal surface of the donor cornea with gentian violet before trephination will help solve this problem by providing a means to ensure correct graft orientation. A correctly placed graft will display a correctly oriented “S” in the recipient eye. If the endothelial surface of the graft is accidentally apposed with the recipient stroma, the “S” will appear backward.
Surgical procedure in the recipient eye. Two side port incisions are made near the three and nine o’clock positions. The anterior chamber is then filled with balanced salt solution, and BSS is used to maintain anterior chamber pressure throughout the procedure. A small superior clear corneal or scleral tunnel incision (approximately 3 to 4 mm) is marked with calipers and made with a keratome blade.
A circle is marked on the recipient corneal epithelium with a 7 to 9 mm trephine densely coated with gentian violet. A bent needle or custom instrument is then used to score the Descemet’s membrane on the posterior side of the cornea along the circle marked on the anterior cornea.
The Descemet’s membrane and endothelium is then stripped using either a custom Descemet’s stripper or an irrigation/aspiration hand piece. Laying the Descemet’s membrane on the anterior surface of the eye after dissection helps to verify that all of the desired tissue has been removed. Injecting trypan blue into the anterior chamber also aids in the visualization of any remaining Descemet’s membrane.
Two drops of viscoelastic material are applied to the endothelial surface of the donor button. We fold the button so that the endothelium is on the inside, with 60 percent of the button on one side of the fold and 40 percent on the other. The folded graft is grasped gently with angled Kelman or IOL forceps and introduced into the recipient anterior chamber through the 3 to 4 mm incision.
Alternatively, the graft may be folded twice (as one would fold a letter) and inserted through a smaller incision. (The long-term effects, however, of graft folding on the viability and function of endothelial cells are not yet known.) Care should be taken not to apply excess pressure to the graft with forceps, as this could also damage the endothelium.
The graft is unfolded by gently sliding a cannula into the fold of the graft and introducing BSS. It is centered using peripheral pressure from the flow of BSS into the chamber. After positioning, the entire recipient stroma should be covered by donor tissue because areas of bare stroma might increase postoperative edema. The wound is closed with 10-0 nylon suture.
Finally, sterile air is delivered to the eye via a cannula through one of the side port incisions to make an air bubble with borders extending past the edges of the graft. This bubble holds the graft in place until it has adhered to the recipient stroma.
Postoperative care. The patient should remain supine for approximately one hour after surgery to lower the risk of graft dislocation. After this amount of time, the Na+, K+-ATPase in the endothelium should create enough suction to fix the graft in position. The eye should be examined with a slit-lamp biomicroscope and the IOP checked one hour after surgery. In order to avoid pupillary block, the pupil should be kept dilated. Air can also be released from the anterior chamber through a side port, at the slit lamp, to clear the inferior pupillary border.
Patients may use a regimen of corticosteroids and antibiotics comparable to that prescribed after traditional PK.
Conclusions
DSAEK is a recently described technique, and the fine-tuning of this procedure is far from complete. The most common complication of DSAEK is graft dislocation shortly after surgery, which results in lowered graft endothelial cell density even after the graft has been repositioned. Dislocations are more common early in the surgical learning curve, and techniques are being identified to lower their incidence.3 Time will reveal how the extended survival of donor endothelial cells in DSAEK compares with that in full-thickness grafting, and these results may determine the fate of this procedure. Further studies are needed to better understand the effects of cutting and folding the cornea on endothelial viability and function. In the meantime, the early results with respect to postoperative astigmatism make DSAEK a promising new alternative to PK.
____________________________
1 Edelhauser, H. F. Invest Ophthalmol Vis Sci 2006;47:1755–1767.
2 Joyce, N. C. Prog Retin Eye Res 2003;22:359–389.
3 Price, F. W. et al. J Cataract Refract Surg 2006;32:411–418.
____________________________
Mark Fernandez is a fourth-year medical student, and Dr. Afshari is an associate professor of ophthalmology. Both are at Duke University. Financial disclosure: Dr. Afshari has a research award from Research to Prevent Blindness.