To evaluate complications and clinical outcomes of Descemet stripping automated endothelial keratoplasty in eyes with preexisting anterior chamber intraocular lenses.
Thirty-one patients who underwent Descemet stripping automated endothelial keratoplasty and who had a preexisting anterior chamber intraocular lens were identified from May 2006 through March 2009. Patient follow-up ranged from a minimum of 1 month up to 30 months. Preoperative and postoperative best spectacle-corrected visual acuity, manifest refraction, comorbid conditions, and complications were recorded. Endothelial cell loss, graft dislocation, graft failure rates were calculated.
The mean age at surgery was 78 ± 9 years (range, 53 to 91 years). All eyes had pseudophakic bullous keratopathy, except 1 patient who had a failed penetrating keratoplasty graft. Excluding those patients with severely limited visual potential because of noncorneal pathologic features, the mean best spectacle-corrected visual acuity improved significantly from 20/200 to 20/400 before surgery to 20/63 at 3 months ( P < .0001), 20/60 at 6 months ( P = .0006), 20/50 at 12 to 15 months (n = 10; P = .004), and 20/40 between 23 and 30 months (n = 8; median, 25 months; P = .007). The preoperative mean spherical equivalent was −0.3 ± 1.8 diopters compared with −0.15 ± 1.5 diopters after surgery ( P = .78). The graft dislocation rate was 13%, and the graft failure rate was 16%. For those patients with endothelial cell density data available, the average endothelial cell loss was 48% at a mean of 14 months.
Although Descemet stripping automated endothelial keratoplasty surgery in patients with an anterior chamber intraocular lens remains a controversial topic, the visual and anatomic outcomes from this limited study support this approach as a surgical option in selected cases.
Corneal endothelial dysfunc tion is a common indication for corneal transplantation. Descemet stripping automated endothelial keratoplasty (DSAEK) is a newer corneal surgery technique in which the diseased host endothelium and Descemet membrane are removed and replaced with donor posterior lamella. This procedure offers faster visual recovery compared with penetrating keratoplasty. Complications of DSAEK include graft dislocation, primary graft failure, and pupillary block. The presence of an anterior chamber intraocular lens (AC IOL) in patients requiring DSAEK has the potential to make this procedure more complicated secondary to decreased anterior chamber depth, difficulty in unfolding the graft, escape of air via an existing peripheral iridectomy, or a combination thereof. There have been reports of an AC IOL exchange for a sutured posterior chamber intraocular lens (PC IOL) in patients both before or at the time of DSAEK. Little has been reported in the literature regarding the outcomes of DSAEK in patients with preexisting AC IOLs who do not undergo intraocular lens exchange. The purpose of this study was to evaluate visual outcomes, refractive changes, and DSAEK complications in the management of corneal endothelial disorders in eyes with pre-existing AC IOLs.
This study was a retrospective review of patients undergoing DSAEK with preexisting AC IOLs from March 2006 through March 2009. Thirty-one eyes of 31 patients undergoing DSAEK with preexisting AC IOLs were identified; all patients had minimum of 1 month of follow-up. All surgeries were performed by 1 of 3 surgeons (T.K., D.T.V., or N.A.A.). The medical chart was reviewed for the following parameters: age; comorbid conditions; preoperative and postoperative refractive error; visual acuity before surgery and at 3, 6, and 12 months or more after surgery; and surgical complications. Data were analyzed using Microsoft Excel (Microsoft, Redmond, Washington, USA); the Wilcoxon matched pairs signed-rank test was performed on all values to determine significance.
Surgery was performed under monitored anesthesia care with topical anesthesia. Paracentesis incisions were made in the supratemporal and inferotemporal quadrants. After infusion of nonpreserved lidocaine hydrochloride 1%, the anterior chamber was inflated with a cohesive viscoelastic. The corneal epithelium was marked with an 8.0-mm trephine and Gentian violet ink marking pen. A bent, 25-gauge needle was inserted through one of the paracentesis incisions and was used to score the Descemet membrane 360 degrees corresponding to the trephine mark. A temporal clear corneal incision then was made in the host with a 2.5-mm keratome. A Gorovoy stripper on a 3-mL syringe then was used to strip off the host endothelium and Descemet membrane. The irrigation–aspiration unit was used to remove any remaining viscoelastic material. The wound was enlarged to approximately 4 mm. Precut DSAEK donor graft tissue was trephinated to the appropriate size (range, 8.0 to 8.5 mm in all cases depending on the host corneal diameter) and then was brought to the operative field. A small amount of viscoelastic was placed on the endothelial surface of the graft, which then was folded in a 60/40 so-called taco fashion. The graft was inserted into the anterior chamber using a Utrata-type forceps. Filtered air was injected into the anterior chamber to unfold the graft and to attach completely the precut DSAEK donor graft to the host cornea. For the occasional instances where the air bubble migrated to the posterior segment, the iris or eye position, or both, were manipulated to bring the air bubble back into the anterior chamber. In the rare instance where graft unfolding was not able to be accomplished by air injection alone, a second instrument (i.e., a Sinkey hook) was inserted through the second paracentesis incision to assist in unfolding of the graft. A 10-0 nylon suture was placed in the operative wound. Finally, 4 venting incisions were placed in the peripheral cornea to release any fluid in the graft–host interface to facilitate graft adherence. After waiting 8 minutes with the patient in the supine position with 100% air fill of the anterior chamber, the patient was undraped, given 1 drop of topical moxifloxacin 0.5% and prednisolone acetate 1%, and taped with a Fox shield on the operative eye. The patients then were instructed to lay supine for 2 hours in the postoperative recovery room, after which they were examined at the slit-lamp biomicrscope to ensure graft attachment and to check intraocular pressure. All cases reviewed had a prior patent peripheral iridectomy in place. For any case in which intraocular pressure was elevated, topical medications were given, and if necessary, a small amount of air was released via the paracentesis incision at the slit lamp. All patients were continued on moxifloxacin 0.5% and prednisolone acetate 1% 4 times daily for the first week after surgery, and prednisolone acetate 1% was continued 4 times daily over the first 1 to 2 months after surgery, and then the drugs were tapered gradually.
Thirty-one charts of patients who underwent DSAEK with a pre-existing AC IOL were identified. The mean age of all patients at surgery was 78 ± 9 years (range, 53 to 91 years). Twenty-nine percent (9/31) were female. All patients had a diagnosis of pseudophakic bullous keratopathy (16% [5/31] of these patients also had various stages of Fuchs dystrophy), except one eye with a failed full-thickness graft (the prior penetrating keratoplasty was performed for keratoconus). Comorbid conditions included: history of primary open-angle glaucoma (n = 14; 3 of these patients had glaucoma tube shunts present before surgery), cystoid macular edema (n = 8), neovascular and nonneovascular age-related macular degeneration (n = 2), retinitis pigmentosa (n = 1), macular hole (n = 1), hypotony maculopathy (n = 1), retinal detachment (n = 1), optic atrophy (n = 1), and keratoconus (n = 1). In total, 24 of the 31 patients had a prior history of 1 or more of the mentioned comorbid conditions. All had mild noncorneal pathologic features except for 7 patients who had history of severe retinal or optic nerve pathologic features, or both (retinal detachment, macular hole, neovascular macular degeneration, hypotony maculopathy, end-stage glaucoma, optic atrophy, or a combination thereof), with preoperative and postoperative acuity ranging from 20/200 to counting fingers. These patients were excluded from the visual acuity analysis because their retinal or optic nerve disease, or both, severely limited best potential visual acuity, despite improvement in corneal endothelial function and corneal clarity. The remaining cases were included despite presence of comorbid conditions because it was deemed that there was potential for significant visual recovery. In all cases, no significant peripheral anterior synechiae were documented to be present in the anterior chamber. Two cases required limited anterior vitrectomy at the time of DSAEK surgery to remove vitreous in the anterior chamber. Although not specified in most of the charts reviewed, most of the AC IOLs were open-loop models on inspection during the operative case.
Mean BSCVA before surgery was 20/200 to 20/400 for all patients included in the visual acuity analysis (n = 24; 7 patients were excluded because of severe noncorneal pathologic features). All patients did not have follow-up data at all time points. Mean BSCVA at postoperative month 3 was 20/63 (n = 17; P < .0001). At 6 months, mean BSCVA was 20/60 (n = 14; P = .0006). At 12 to 15 months, the mean BSCVA was 20/50 (n = 10; P = .004). Eight patients had follow-up between 23 and 30 months (median, 25 months). The mean BSCVA in this group was 20/40 ( P = .007). Both preoperative and postoperative manifest refraction data were available for 16 patients. The postoperative cylinder of 2.1 ± 1.7 diopters (D) did not differ significantly from 2.6 ± 1.7 D before surgery ( P = .46). Preoperative mean spherical equivalent was −0.3 ± 1.8 D, compared with −0.15 ± 1.5 D after surgery ( P = .78).
Complications reported for all patients in this study included 5 patients with graft failure (5/31; 16%), 4 graft dislocations (4/31; 13%) requiring rebubbling, 1 transient hyphema (1/31; 3%), and no cases of graft rejection or pupillary block. All graft dislocations occurred between postoperative days 1 and 7 and were refloated successfully. One patient with graft dislocation also had a glaucoma tube shunt present before surgery. There were no cases of immediate graft detachment noted at the time of slit-lamp examination on postoperative day 0. Of the graft failure cases, 3 were primary failures and 2 occurred late in the postoperative course (between 5 and 12 months). Three of the 5 patients with graft failure had pre-existing glaucoma, 2 of whom had glaucoma tube shunts present at the time of surgery (1 patient in the primary failure group and 1 patient in the late failure group). Two of the 5 graft failure patients also had history of cystoid macular edema. No patient in the study had multiple complications (i.e., patients who had graft dislocation did not also have graft failure).
Preoperative donor graft endothelial cell count (ECC) was available in 27 of 31 patients, and the mean ECC before surgery was 3234 cells/mm 2 . Comparative preoperative and postoperative ECC data were available in 10 patients, of whom 5 patients had experienced postoperative complications (3 successfully managed graft detachments and 2 late graft failures). The overall average endothelial cell loss was 48% at a mean of 14 months (range, 3 to 34 months; median, 15 months). When divided by the presence or absence of a postoperative complication, the average ECC was 1853 cells/mm 2 (43% loss) in the uncomplicated patients, compared with 1449 cells/mm 2 (59% loss) in the complicated group, although this was not statistically significant ( P = .37).