Corneal Graft Folds: A Complication of Deep Lamellar Endothelial Keratoplasty | Cornea | JAMA Ophthalmology | JAMA Network
[Skip to Navigation]
Sign In
Clinicopathologic Reports, Case Reports, and Small Case Series
April 2006

Corneal Graft Folds: A Complication of Deep Lamellar Endothelial Keratoplasty

Arch Ophthalmol. 2006;124(4):593-595. doi:10.1001/archopht.124.4.593

Corneal endothelial dysfunction accounted for 36% of the more than 30 000 corneal transplants performed in the United States in 2003.1 Deep lamellar endothelial keratoplasty (DLEK) is an evolving procedure2-4 that, in comparison with full-thickness keratoplasty, may result in a decreased incidence of high or irregular astigmatism and suture-related complications, such as microbial keratitis or sterile suture reactions. Published complications associated with DLEK have included graft rejection, mismatch of donor thickness to recipient bed thickness, and partial dislocation of the graft.3,4 We report another complication encountered in 2 patients undergoing DLEK.


Two patients were randomly assigned to DLEK as part of an institutional review board–approved, prospective, randomized study comparing DLEK and standard penetrating keratoplasty. Deep lamellar endothelial keratoplasty was performed through a 9-mm limbal incision in the manner described by Terry and Ousley.3 An 8-mm posterior lamellar disc of recipient tissue was excised and replaced with a partial-thickness donor disc 8 mm in diameter. Patient 1 also underwent concomitant phacoemulsification and intraocular lens implantation through a separate temporal, clear corneal wound. Intraoperatively, the donor disc was noted to be very thin. The procedure was otherwise uncomplicated. Patient 2, a 74-year-old woman who was previously pseudophakic, underwent an apparently uncomplicated DLEK for Fuchs dystrophy as well. Postoperatively, potential visual acuity was measured (Guyton/Minkowski PAM; Mentor, Inc, Norwell, Mass) and corneal thickness was calculated using confocal microscopy (Tandem Scanning Microscope, Reston, Va).5


Patient 1: On the first postoperative day, the graft was in good apposition to the host tissue, but vertical folds were noted in the center of the donor tissue. Corrected visual acuity was 20/80-1. At 6 months, corrected visual acuity was 20/70 with central folds persisting (Figure 1). Distortion was noted on retinoscopy and in the red reflex during slitlamp examination. The potential acuity was 20/60. The fundus examination results were normal, and there was no evidence of macular edema by optical coherence tomography (Stratus 3000, Carl Zeiss Meditec, Dublin, Calif). The thickness of the donor tissue was 42 μm with an overall corneal thickness of 359 μm. Replacement of the donor tissue was offered, but the patient elected against further surgery.

Figure 1. 
Case 1. Folds in the lamellar graft are visible on retroillumination.

Case 1. Folds in the lamellar graft are visible on retroillumination.

Patient 2: On the first postoperative day, the graft was not apposed in 1 quadrant and was twice repositioned on postoperative days 6 and 7 with additional air injections. As the graft thinned, large, vertical Descemet folds were noted on the 12th postoperative day centrally with visual acuity, without correction, at 20/400. At 6 months, the folds were still present (Figure 2), and corrected visual acuity was 20/125 with a measured potential visual acuity of 20/70. Again, distortion was noted on retinoscopy as well as in the red reflex. The thickness of the donor tissue was 108 μm with an overall corneal thickness of 564 μm. A successful repeat DLEK was then performed by replacing the graft. Pathologic examination of the surgical specimen showed a full-thickness fold in the graft (Figure 3).

Figure 2. 
Case 2. Folds are present in the center of the graft (A and B).

Case 2. Folds are present in the center of the graft (A and B).

Figure 3. 
Case 2. Toluidine blue–stained specimen of the excised donor graft demonstrating a fixed fold.

Case 2. Toluidine blue–stained specimen of the excised donor graft demonstrating a fixed fold.


In the management of corneal endothelial disorders, DLEK may be a reasonable alternative to penetrating keratoplasty, considering the advantage of more predictable refractive results. However, 1 potential drawback of DLEK is opacity or optical irregularity of the donor-host interface, which may limit the postoperative vision.

Other than interface-related concerns, we are unaware of any previous reports describing complications that affect the optical quality of the graft. We feel that the fixed folds of the donor graft contributed directly to an adverse visual outcome in our 2 patients. In both cases, the measured potential acuity was better than the corrected visual acuity. The graft tissue and endothelium appeared healthy without other abnormalities to suggest an additional cause of poor optical performance. Regarding the cause of this complication, we speculate that the thin donor lamella in patient 1 may not have provided adequate stromal support to prevent wrinkling of the graft tissue. We do not know whether a larger or longer-acting gas bubble with prolonged supine positioning would have prevented the occurrence of folds or would have remedied the problem if attempted in the early postoperative period.

Possible causes for the folds with patient 2 are more difficult to identify. This donor graft was not in adequate apposition to the host tissue initially and required repositioning twice. A delay in creating complete donor-host adherence may have predisposed to these folds. Perhaps compression of the graft tissue may result from improper centration within the recipient bed or an inadvertently undersized recipient bed diameter. Manual dissection of the recipient bed may also lead to irregularities that do not allow for maximal donor-host adherence and positioning.

These 2 cases illustrate a complication of DLEK that we feel led to an undesirable visual outcome. Considering the potential causes for this condition, we feel that there is currently a lack of understanding of the ideal donor thickness and donor and host diameters. If irregularities secondary to the manual dissection of the corneal tissues, with resulting disparity in donor-host shape or diameters, are demonstrated to be a factor in future cases, one might argue for intentionally oversizing the recipient bed diameter. Alternatively, automated dissection, such as laser-assisted techniques, may ensure ideal lamellar thickness and size. We also query whether early treatment of future patients with these folds, whether by prolonged gas tamponade or repositioning of the donor button, might improve the postoperative optical performance of the cornea graft.

Correspondence: Dr Baratz, Department of Ophthalmology, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (

Submitted for Publication: January 4, 2005; final revision received February 8, 2005; accepted February 15, 2005.

Financial Disclosure: None.

Funding/Support: This study was supported by grant EY02037 from the National Eye Institute, Bethesda, Md, and Research to Prevent Blindness, Inc, New York, NY.

Acknowledgment: We thank Douglas J. Cameron, MD; Thomas P. Link; and Jay A. Rostvold for histopathologic and photographic support.

Eye Bank Association of America, Statistical report. 2003;
Melles  GRLander  FBeekhuis  WHRemeijer  LBinder  PS Posterior lamellar keratoplasty for a case of pseudophakic bullous keratopathy.  Am J Ophthalmol 1999;127340- 341PubMedGoogle ScholarCrossref
Terry  MAOusley  PJ Replacing the endothelium without corneal surface incisions or sutures: the first United States clinical series using the deep lamellar endothelial keratoplasty procedure.  Ophthalmology 2003;110755- 764PubMedGoogle ScholarCrossref
Terry  MAOusley  PJ Rapid visual rehabilitation after endothelial transplants with deep lamellar endothelial keratoplasty.  Cornea 2004;23143- 153PubMedGoogle ScholarCrossref
McLaren  JWNau  CBErie  JCBourne  WM Corneal thickness measurement by confocal microscopy, ultrasound, and scanning slit methods.  Am J Ophthalmol 2004;1371011- 1020PubMedGoogle ScholarCrossref