• Ophthalmic Pearls

    Deep Anterior Lamellar Keratoplasty

    By Vishak John, MD, Kenneth M. Goins, MD, and Natalie A. Afshari, MD
    Edited by Sharon Fekrat, MD, Ingrid U. Scott, MD, MPH, and Woodford S. Van Meter, MD

    Deep anterior lamellar keratoplasty (DALK) is a surgical procedure for removing the corneal stroma down to Descemet’s membrane. It is most useful for the treatment of corneal disease in the setting of a normally functioning endothelium. Traditionally, penetrating keratoplasty (PK), which involves a full-thickness corneal graft, has been the treatment of choice for corneal stromal diseases. But PK can be complicated by graft rejection, irregular astigmatism and corneal opacification, thus resulting in visual impairment. DALK offers an alternative procedure that may lessen those risks because the recipient Descemet’s membrane and endothelium are preserved. At the same time, DALK carries the potential danger of decreased visual acuity due to possible opacification at the interface layers.

    Brief History

    As early as the 1950s, Jose Barraquer and colleagues in Colombia applied new techniques of lamellar keratoplasty, dissecting the corneal stroma down to two-thirds of its thickness in both the donor and the recipient tissue.1 Yet the procedure failed to gain favor because of poor visual outcomes related to irregularity of the dissected surfaces and scarring in the tissue interfaces.2 Although exposure of Descemet’s membrane in corneal dissection was performed in the 1970s, the term “deep lamellar keratoplasty,” as it is used today, was not employed until 1984 by Eduardo Arenas Archila, MD, with the use of intrastromal air injection to facilitate host tissue removal.3 By the late 1990s, studies indicated that DALK was associated with visual outcomes similar to PK without the risk of immunological rejection.2 In spite of positive reports in the literature, the classic technique of layer-by-layer stromal tissue removal was tedious and required great surgical experience, thereby limiting its use around the world. Recent advances in techniques contributed by many surgeons have started to popularize the procedure.


    DALK can be an effective treatment for any pathology of the anterior cornea (epithelium, Bowman’s layer and stroma) as long as the patient has an intact, functioning endothelium. Common indications for DALK include keratoconus and corneal scars. Patients with keratoconus are good candidates for DALK because they are typically young and have healthy endothelium. These patients stand to lose the most from the occurrence of post-PK immunological reactions that can compromise endothelial function in up to 20 percent of cases. Less common indications for DALK include vernal keratoconjunctivitis, corneal dystrophies and ocular surface diseases with limbal stem cell deficiency, including Stevens-Johnson syndrome, ocular cicatricial pemphigoid and chemical/thermal burns.3

    Pros and Cons

    DALK offers a variety of theoretical and practical advantages compared with PK for patients with a healthy corneal endothelium.

    Pros: Anatomic outcomes. Eighteen percent of primary full-thickness grafts have been reported to fail within 10 years due to endothelial rejection and chronic endothelial cell loss,2 and DALK may reduce this rejection rate because the host endothelium is preserved.2,4 In a randomized clinical trial comparing DALK and PK published in 2002, there was a significant difference in endothelial density between treatment groups at 24 months postoperatively (2,183 ± 300 vs. 1,868 ± 272 per mm2 respectively, P = 0.044).4 Furthermore, while endothelial cell density stabilized around six months after DALK, there continued to be loss of cells in the PK group at 24 months.4 Moreover, keratometric astigmatism rates were lower at six and 12 months after DALK compared with the same intervals after PK.5 Additionally, studies indicate a shorter healing time (suture removal at four months after DALK vs. 12 months with PK),2 fewer postoperative complications and shorter use of topical steroid treatment with the deep anterior approach. Finally, from a public health perspective, DALK provides the opportunity to use donor tissue with questionable endothelium that might not be suitable for full-thickness grafts. In the future, it may be possible to use the same donor cornea for a DALK and a Descemet’s stripping endothelial keratoplasty.

    Visual outcomes. In spite of the clinical and theoretical advantages of lamellar keratoplasty, visual outcomes have not always been ideal. Visual acuities associated with older lamellar keratoplasty techniques were inferior to those associated with PK due to the tissue interface scarring that developed between the donor cornea and the recipient stromal bed. However, since deep lamellar keratoplasty surgery removes almost all of the recipient stromal tissue, the risk of scarring in the interface and subsequent poor visual outcomes is much reduced.

    Uncertain. Currently, the literature is ambivalent regarding visual outcomes between DALK and PK. In one comparative cohort study, the proportion of patients who achieved a BCVA of 20/20 was significantly higher following PK compared with DALK (70 percent vs. 22 percent, P = 0.04);5 however, in another randomized, controlled trial, there was no statistically significant difference between the two procedures in associated postoperative BCVA.4 Ultimately, the visual result of DALK depends on the ability to maintain clear interfaces between the tissues. The femtosecond laser has been successfully used in donor human eyes to cut posterior lamellar flaps in endothelial keratoplasty, and this technology could potentially reduce tissue irregularity and scarring in DALK.

    Cons. The disadvantages of the procedure are the complexity and novelty of the techniques. A learning curve exists, and cases of recipient corneal perforation requiring transition to PK may require another donor eye with better endothelial quality.

    Performing DALK

    Various approaches for performing DALK have been described in the literature. One method is to remove the host anterior corneal tissues layer-by-layer until reaching the deep stroma or the bare Descemet’s membrane. Although different surgical techniques vary in their details, the basic surgery consists of the following steps:

    Recipient eye. The anterior corneal surface is cut with a suction trephine set to a depth of about two-thirds of the corneal thickness. Then the stromal layers are dissected with a rounded blade, angled parallel to Descemet’s membrane. Fluid or air is then injected using either a 27- or 30-gauge cannula in between the deep stroma and Descemet’s membrane to separate those layers.

    Because early techniques failed to visualize the depth of stromal dissection, there was a greater risk of perforation than when Descemet’s membrane can be visualized. New techniques described below can decrease surgical times while improving the safety and success rates of DALK:

    • Intrastromal air injection. One innovation in DALK has been the use of an air-filled tuberculin syringe, the needle of which is injected obliquely into the stroma prior to trephination. This intrastromal air renders the cornea opaque and provides a safe deep interface for the trephine.1
    • Hydrodelamination. After the initial trephination to about 75 percent of the corneal thickness, the surgeon can inject balanced salt solution through a cannula into a small pocket created in the central stroma. The saline induces stromal fiber swelling, which facilitates fine manipulation with forceps.3
    • Viscoelastic dissection. After an initial 80 to 90 percent trephination, sodium hyaluoronate can be injected through a blunt cannula deep into the central corneal lamella near Descemet’s membrane. Injection of the viscoelastic substance between the deep stroma and Descemet’s membrane facilitates the separation of the final layers.6
    • Big bubble. With the “big bubble” technique, air is injected deep into a groove created by trephining 60 to 80 percent stromal thickness.1 This introduction of air into the stroma anterior to Descemet’s membrane creates a dome-shaped detachment of Descemet’s membrane, which is then identified by a ring visible with the microscope.3
    • Anterior chamber air. In order to obtain the best visualization during the surgery, Gerrit R. J. Melles, MD, PhD, suggested introduction of air in the anterior chamber. This injected air creates a mirrorlike effect that facilitates the movement of surgical instruments between Descemet’s membrane and the deep stroma. Furthermore, the air-to-endothelium interface becomes a landmark to identify the posterior surface of the cornea, serving as a reference for dissection depth.3

    Donor eye. Although preserved grafts can be used, most of the literature reports use of fresh corneas prepared by the surgeon. Descemet’s membrane and endothelium are removed by gently swabbing the posterior corneal surface of the donor corneoscleral rims with dry cellulose sponges. Forceps also may be used for removing the posterior corneal layers. Then a corneal button is punched out from the tissue. Suturing technique (interrupted, running or combined interrupted-running 10.0 nylon sutures) can be done according to the surgeon’s preference in DALK, just as in PK. After suturing, a bandage soft contact lens is placed on the cornea.


    The most frequently encountered complication of DALK is perforation of Descemet’s membrane and entering the anterior chamber from the stroma. Descemet’s membrane and the posterior cornea are especially susceptible to perforation by sharp instruments. Tears or perforations occur in approximately 10 to 30 percent of cases.2 The viscoelastic dissection technique may offer some protection against perforation. If a perforation occurs, management depends on the timing and the size of the injury. If the tear occurs while the stroma still covers Descemet’s membrane, it generally self-seals. Small perforations occurring during dissection may result in postoperative Descemet’s membrane detachment. Most Descemet’s membrane detachments can be managed by injecting air or an air/gas mixture (sulfur hexafluoride + air) into the anterior chamber after surgery or under the slit lamp postoperatively, thereby creating a tamponade to seal the Descemet’s membrane defect.1,3 However, it should be noted that the use of a tamponade to reattach Descemet’s membrane may lead to closure of the anterior chamber angle from pupillary block. A rare complication occurs when a fixed, dilated pupil ensues secondary to iris ischemia, referred to as the Urrets-Zavalia syndrome.7 In cases where the tamponade fails to manage the perforation, a conversion to traditional penetrating keratoplasty might be necessary.

    Postoperative Care

    Topical corticosteroids and antibiotics are administered postoperatively. Patients are monitored for inflammation, infection, graft rejection, astigmatism and suture-related problems. Overall, patients tend to fare well with a shorter course of corticosteroids after DALK than after PK. At 12 months postoperatively in one study, only 15 percent of DALK patients used topical prednisolone, while 75 percent of PK patients continued to need the treatment.3 In addition, sutures can be generally removed earlier in DALK patients than in PK patients.


    With a long history, DALK is seeing a recent renewed interest. Although DALK has the reputation of being challenging and time-consuming, recent advances in techniques have cut the surgical time considerably. In patients with anterior corneal opacification or structural corneal defects with a healthy, functioning endothelium, DALK can be considered as the first-line treatment due to its preservation of host endothelium and Descemet’s membrane. If future studies show that DALK provides lower postoperative complications, fewer graft rejections and similar visual outcomes compared to PK, DALK should have a promising future in corneal surgery as a viable alternative to full-thickness PK in selected patients.


    1 Anwar, M. and K. Teichmann. Cornea 2002;21(4):374–383.

    2 Trimarchi, F. et al. Ophthalmologica 2001;215:389–393.

    3 Shimmura, S. and K. Tsubota. Current Opin Ophthalmol 2006;17:349–355.

    4 Shimazaki, J. et al. Am J Ophthalmol 2002;134:159–165.

    5 Funnel, C. L. et al. Eye 2006;20:527–532.

    6 Manche, E. et al. Arch Ophthalmol 1999;117:1561–1565.

    7 Maurino, V. et al. Am J Ophthlamol 2002;133:266–268.


    Dr. John will be an ophthalmology resident at Emory University. Dr. Goins is associate professor of clinical ophthalmology at the University of Iowa. Dr. Afshari is associate professor of ophthalmology at Duke University.