Retrocorneal membrane following penetrating keratoplasty (PKP) is a known rare complication, which may occur either as an inflammatory membrane1 or as a result of fibrous downgrowth2 or unintentional or intentional retention of host Descemet’s membrane (DM).3–14
Longstanding severe stromal edema has been suggested to lead to loosening of the attachment of host DM, thereby predisposing to its easy separation from the overlying edematous stroma, a factor that contributes to inadvertent incomplete trephination of edematous corneas such as in pseudophakic bullous keratopathy, leading to unintentional retained host DM following PKP.3,4,14
Lowenstein et al.6 published their technique of intentional retention of the host DM in high risk eyes, where opening of the anterior chamber might be associated with severe intraoperative complications during PKP.
The inadvertent retained host DM found rarely following PKP may be associated with progressive opacification, reducing the patient’s visual acuity, or it may threaten the health of the corneal transplant, due to direct contact and interference with endothelial cell metabolism. These two major problems necessitate either Nd:YAG laser optical opening7,9,10 or surgical removal8 and surgical excision.3,13
According to previous reports, aqueous humor flow inside the supernumerary anterior chamber or double anterior chamber, created between the donor corneal graft endothelium and the retained host DM, allows the corneal graft to remain clear for years.1,3–6 However, this vital flow of aqueous requires one or more openings in the retained DM itself or a passage between its border and the peripheral host cornea, as noted by authors who performed a microperforation by a 25-gauge needle at the retained host DM periphery before suturing the corneal graft.5,6 At times, this microperforation seals, thereby necessitating Nd:YAG laser membranotomy to be performed6 to renew the aqueous flow into the supernumerary anterior chamber, maintaining the normal metabolism of the corneal graft’s endothelial cells. This reopening also prevents collapse of the supernumerary chamber with consequent contact between retained DM and the graft endothelium, which may lead to cellular damage.7
We describe five patients with longstanding pseudophakic bullous keratopathy, in whom inadvertent retained DM was observed following PKP. This membrane gradually opacified, reducing the patients’ visual acuity, and therefore Nd:YAG laser central membranotomy was performed. The graft failed within a relatively short time following this intervention.
Design and Methods
In a retrospective review of 1,350 PKPs performed at Rabin Medical Center between 1986 and 2008, we found five patients (0.37%) who presented with inadvertent retained host DM 24 hours following surgery. The patients’ ages ranged between 72 and 85 years and all had longstanding pseudophakic bullous keratopathy or corneal graft failure with marked corneal stromal and epithelial edema, causing low visual acuity that ranged between counting fingers and hand motions.
All patients underwent cataract surgery 10 to 17 years before current admission for PKP, and a posterior chamber intraocular lens was implanted during their surgery. In two patients the intraocular lens was implanted in the capsular bag and in three patients the intraocular lens was implanted in the ciliary sulcus, due to capsular tear with vitreous loss requiring anterior vitrectomy. Two of the latter patients underwent PKP 8 to 10 years before their current repeat keratoplasty and subsequently their corneal graft failed, leading to severe bullous keratopathy; they were therefore scheduled for a second corneal graft. The two patients with pseudophakic bullous keratopathy and intraocular lens implanted in the capsular bag had Fuch’s endothelial dystrophy, diagnosed several years before their cataract surgery.
The five patients underwent uneventful PKP, with an 8.5 mm donor cornea, sutured to an 8.0 mm recipient bed by 10-0 Ethilon sutures (Ethicon, Inc., Somerville, NJ).
On the first day following surgery, a glistening paper thin, transparent, retained host DM was noticed at an estimated depth of 1.5 to 2.0 mm behind the corneal transplant, within a supernumerary anterior chamber. We have also noticed one or two slit openings (dehiscences) in the retained DM periphery, between its border and the host peripheral cornea, close to the graft–host interface, where the supernumerary chamber was found to be shallow. One month following corneal transplantation, best-corrected visual acuity (BCVA) of these patients ranged between 6/20 and 6/30(−). Subsequently, during a period of 3 to 4 months, the host DM gradually opacified and markedly reduced the patients’ visual acuity. However, their corneal graft remained clear with normal endothelial cell count (2,200 to 2,400 cells/mm2) and normal cell morphology. At 4 to 5 months following PKP, the patients’ BCVA deteriorated and was found to be between 6/40(−) and 6/90, due to opacification of the retained DM.
Therefore, Nd:YAG laser (Coherent Inc., Santa Clara, CA) central membranotomy was performed using the Abraham Capsulotomy positive lens with relatively low energy levels (1.4 to 1.6 mJ) that were adjusted according to the thickness of this DM and the fewest possible number of pulses (8 to 10 pulses) to make a diamond-shaped opening of 2.0 to 2.5 mm horizontal diameter at the center of the retrograft host DM.
We took appropriate measures to avoid a peak of intraocular pressure elevation following the YAG laser membranotomy by treating the patients with antiglaucoma medications 2 to 3 hours before and during the first 24 hours following the procedure, including Cosopt (Merck, Whitehouse Station, NJ) and Iopidine (0.5%). All five patients were treated regularly by local steroids, four times daily, before and following the procedure.
At the 2-week examination following YAG laser membranotomy, the BCVA of all five patients improved from a range of 6/40(−) to 6/90 to a range of 6/15(−) to 6/20. The intraocular pressure was normal, ranging between 15 and 20 mm Hg, without antiglaucoma medications. The corneal graft of all patients was clear and without any evidence of corneal edema or Descemet folds and all had normal pachymetry ranging between 500 and 530 microns. Unfortunately, specular endothelial microscopy was not performed following the YAG laser procedure. The central diamond-shaped 2.0 to 2.5 mm horizontal diameter opening enlarged to a trapezoid-shaped aperture of 3.5 to 4.0 mm horizontal diameter, due to peripheral fibrous radial traction. The peripheral dehiscence between this membrane and the graft–host interface also enlarged due to the same mechanism of traction (Figure 1).
Figure 1. A photograph of the corneal graft of patient 1, showing the retained host Descemet’s membrane with a trapezoid-shaped central aperture 2 weeks following YAG laser descemetotomy. Note an upper temporal peripheral dehiscence between this membrane and the graft–host interface.
Subsequently, during a short follow-up period of 6 to 8 weeks, the corneal graft of all patients became edematous with the appearance of Descemet folds, progressing to advanced bullous keratopathy and final graft failure within 6 months, despite continued intensive local steroid treatment in addition to NaCl hyperosmotic drops instilled three times daily. The patients’ graft failure resulted in a low visual acuity, ranging between counting fingers and hand motions.
All five patients subsequently underwent repeat uneventful PKP with total removal of the retained host DM during surgery, within a period of 1 year.
Patient data and postoperative outcomes are presented in Table 1.
Table 1: Patient and Nd:YAG Laser Membranotomy Data
Inadvertent retention of DM is a rare complication of PKP that was previously described.3–14 Brown et al.3 were the first to publish three cases in which the recipient’s DM was found behind the corneal graft 24 hours following PKP. They performed surgical dissection of this retrograft membrane in two cases.
The retained DM may gradually opacify from approximately 3 months postoperatively.5,10,14 This process of progressive opacification has been postulated to occur due to fibroblastic activity, depending on the residual stromal tissue retained along with DM in these cases.13
Alternatively, the retained DM can compromise the donor corneal graft endothelium by cellular contact damage or by limiting aqueous flow and diffusion of aqueous humor metabolic nutrients.3,14,15 In most cases, the corneal graft remains clear if no contact has occurred between the retained membrane and the graft endothelial surface.7–10
Nd:YAG laser membranotomy has been described by Archila et al.7 and others6,9,10 to restore vision due to retained DM opacification. Another treatment option is surgical removal or extraction of the retained DM.6,9 Thyagaraijan et al.13 had to perform surgical membranotomy following failed attempts to create a central aperture by YAG laser membranotomy, probably due to marked fibrous thickening of this membrane.
Lazar et al.5 and Lowenstein et al.6 reported that their corneal grafts remained clear between 28 and 36 months of follow-up, provided that contact between the graft endothelium and retained DM is avoided by allowing aqueous flow between them. This was accomplished by performing a peripheral microperforation of the retained DM during surgery and maintenance of a patent channel with YAG laser repeat membranotomy to ensure the presence of a supernumerary anterior chamber with normal aqueous humor flow inside the chamber.5,6
We have performed YAG laser central membranotomy at 4 to 5 months following corneal transplantation in our 5 patients with retained DM due to gradual opacification of this retrograft membrane, which reduced the patients’ visual acuity. Despite the fact that their corneal graft endothelial density was normal (2,200 to 2,400 cells/mm2), with normal endothelial cell morphology, and the fact that the YAG laser energy we used was relatively low (1.4 to 1.6 mJ) with a few laser shots needed to perform a central small opening in DM, the corneal graft became progressively edematous and failed within 6 to 8 weeks following the YAG laser procedure.
In all five patients, there was no contact between the corneal graft endothelium and the retained DM and a relatively shallow supernumerary anterior chamber was present between the two, allowing enough flow of aqueous humor for maintaining the normal metabolism of the graft endothelial cells. Additionally, the YAG laser treatment was performed in the early stage of opacification when DM was still thin enough to be opened by relatively few shots of YAG laser, with low energy levels. However, one should bear in mind that this retrograft membrane was located close to the corneal graft endothelium.
Despite the fact that this study included a limited number of patients, we conclude that Nd:YAG laser membranotomy for retained DM opacification may lead to accelerated corneal graft endothelial decompensation and permanent failure, because it occurred within a relatively short time following YAG laser treatment. The trauma inflicted by the shockwaves of the Nd:YAG laser pulses could have led to endothelial cell damage with subsequent decompensation of the corneal graft of our patients. It should be remembered that the distance between the retained DM and the corneal graft of our patients was relatively short in comparison to the distance of the posterior capsule or iris tissue when performing posterior capsulotomy following cataract surgery or preventive iridotomy. Therefore, in our patients, the mechanical damage from Nd:YAG laser shockwaves was relatively higher, mainly within a relatively closed chamber such as the supernumerary chamber between the retained DM and the corneal graft.
This situation may be similar to performing Nd:YAG laser iridotomy in a case of iris bombe due to pupillary block glaucoma, where the iris surface is close to the corneal endothelial surface and therefore the YAG laser shockwaves may damage the corneal endothelial cells, leading at times to localized or diffuse corneal edema.16–19
Endothelial cell damage following YAG laser iridotomy and capsulotomy has been previously shown in several studies.16–22 One of those studies has shown significant endothelial cell damage induced by YAG laser pulses when they were focused within a distance of 1.0 mm or less behind the corneal endothelial layer.21
Meyer et al.21 found that none of their YAG laser-treated rabbit eyes, undergoing anterior capsulotomy, showed significant endothelial damage. However, the rabbits’ corneal endothelium was focally denuded when laser shots were applied within a distance of 1.0 mm of this layer, and DM was disrupted by laser shots focused within a distance of 0.1 mm. A previous report21 concluded that application of Nd:YAG laser pulses within 1.0 mm of the corneal endothelium should be avoided because significant endothelial damage is likely to occur within this range and recommended that the minimal effective laser energy should be selected when YAG laser application in the vicinity of the endothelium is necessary.
The location of retained DM was within an estimated distance of 1.5 to 2.0 mm from the corneal graft endothelium in our five patients. Therefore, we tried to use the minimal effective energy of our Nd:YAG laser system and the minimal number of shots needed to create a small opening in the central zone of this retrograft membrane. Nevertheless, this proximity of laser shots within the relatively closed supernumerary chamber probably caused significant shockwave damage to the corneal graft endothelium that acted as a stimulus for an accelerated loss of endothelial cells, leading to corneal graft failure.
According to our high rate of accelerated corneal graft endothelial decompensation following Nd:YAG laser membranotomy, we suggest avoiding this procedure in a situation where there is a close proximity of the retained DM to the graft endothelium, and instead performing partial surgical excision of the DM under viscoelastics. We are of the opinion that this is a safer alternative approach to YAG laser treatment, aimed to create a central aperture in the opacified retained host DM.
Future clinical studies are needed to compare the survival of the corneal graft following these two treatment options.
- Lifshitz T, Oshri T, Resenthal G. Retrocorneal membrane after penetrating keratoplasty. Ophthalmic Surg Lasers. 2001;32:159–161.
- Sherrard ES, Ryoroft PV. Retrocorneal membranes: their origin and structure. Br J Ophthalmol. 1967;51:379–386. doi:10.1136/bjo.51.6.379 [CrossRef]
- Brown SI, Dohlman CH, Boruchoff SA. Dislocation of Descemet’s membrane during keratoplasty. Am J Ophthalmol. 1965;60:43–45.
- Hagedoorm A. Corneal transplantation in advanced Fuch’s dystrophy: the fate of homotransplanted human endothelial cells. Am J Ophthalmol. 1967;64:433–436.
- Lazar M, Loewenstein A, Geyer O. Intentional retention of Descemet’s membrane during keratoplasty. Acta Ophthalmologica (Copenh). 1991;69:111–112. doi:10.1111/j.1755-3768.1991.tb02005.x [CrossRef]
- Loewenstein A, Geyer O, Lazar M. Intentional retention of Descemet’s membrane in keratoplasty for the surgical treatment of bullous keratopathy. Acta Ophthalmologica (Copenh). 1993;71:280–282. doi:10.1111/j.1755-3768.1993.tb05005.x [CrossRef]
- Archila AE, Ramirez Cabrera MF, Mieth Alviar A. Double Descemet’s membrane in penetrating keratoplasty. Refract Corneal Surg. 1993;9:65–66.
- Chen YP, Lai PC, Chen PY, Lin KK, Hsiao CH. Retained Descemet’s membrane after penetrating keratoplasty. J Cataract Refract Surg. 2003;29:1842–1844. doi:10.1016/S0886-3350(03)00235-9 [CrossRef]
- Steinemann TL, Henry K, Brown MF. Nd:YAG laser treatment of retained Descemet’s membrane after penetrating keratoplasty. Ophthalmic Surg. 1995;26:80–81.
- Masket S, Tennen DG. Neodymium:YAG laser optical opening for retained Descemet’s membrane after penetrating keratoplasty. J Cataract Refract Surg. 1996;22:139–141.
- Sinha R, Vajpayee RB, Sharma N, Titiyal JS, Tandom R. Trypan blue assisted descemetorhexis for inadvertently retained Descemet’s membrane after penetrating keratoplasty. Br J Ophthalmol. 2003;87:654–655. doi:10.1136/bjo.87.5.654 [CrossRef]
- Ide T, Yoo SH, Kymionis GD, Shah PA, O’Brien TP. Double Descemet’s membrane after penetrating keratoplasty with anterior segment optical coherence tomography. Ophthalmic Surg Lasers Imaging. 2008;39:422–425. doi:10.3928/15428877-20080901-09 [CrossRef]
- Thyagaraijan S, Mearza AA, Falcon MG. Inadvertent retention of Descemet membrane in penetrating keratoplasty. Cornea. 2006;25:748–749. doi:10.1097/01.ico.0000208823.78355.57 [CrossRef]
- Henderson JW, Wolter RJ. Sepepration of Descemet’s membrane in keratoplasty. Am J Ophthalmol. 1968;65:375–378.
- Waring GO 3rd, Bourne WM, Edelhanser HF, Kenyon KR. The corneal endothelium: normal and pathologic structure and function. Ophthalmology. 1982;89:531–590.
- Power WJ, Collum LM. Electron microscopic appearances of human corneal endothelium following Nd:YAG laser iridotomy. Ophthalmic Surg. 1992;23:347–350.
- Wu SC, Jeng S, Huang SC, Lin SM. Corneal endothelial damage after neodymium: YAG laser iridotomy. Ophthalmic Surg Lasers. 2000;31:411–416.
- Kozobolis VP, Detorakis ET, Vlachonikolis IG, Pallikaris IG. Endothelial corneal damage after neodymium:YAG laser treatment: pupillary membranectomies, iridotomies, capsulotomies. Ophthalmic Surg Lasers. 1998;29:793–802.
- Landers J, Craig J. Decompression retinopathy and corneal edema following Nd:YAG laser peripheral iridotomy. Clin Experiment Ophthalmol. 2006;34:182–184. doi:10.1111/j.1442-9071.2006.01180.x [CrossRef]
- Vaikoussis E, Bisogiannis Z, Margaritis L. Corneal endothelial damage after Nd:YAG laser anterior capsulotomy: an experimental study on rabbits. Doc Ophthalmol. 1993;83:279–286. doi:10.1007/BF01204329 [CrossRef]
- Meyer KT, Pettit TH, Straatsma BR. Corneal endothelial damage with Neodymium:YAG laser. Ophthalmology. 1984;91:1022–1028.
- Ker Muir MG, Sherrard ES. Damage to the corneal endothelium during Nd:YAG photodisruption. Br J Ophthalmol. 1985;69:77–85. doi:10.1136/bjo.69.2.77 [CrossRef]
Patient and Nd:YAG Laser Membranotomy Data
|Patient||Age/Sex||VA Before PKP||BCVA 4 to 5 Months After PKP||Timing of Laser Treatment After PKP (Mo)||BCVA 2 Weeks After Laser Treatment||Timing of Graft Failure After Laser Treatment (Wk)||Pulse Energy (mJ)||No. of Pulses|