Descemet Stripping Automated Endothelial Keratoplasty With a Donor Insertion Device: Clinical Results and Complications in 100 Eyes


To study the clinical outcomes of Descemet stripping automated endothelial keratoplasty (DSAEK) with the EndoGlide donor insertion device.


Retrospective interventional case series.


We included 100 eyes that underwent DSAEK for Fuchs endothelial dystrophy or pseudophakic bullous keratopathy (PBK) at a single institution. Eyes with anterior segment pathology or previous intraocular surgery (except for uncomplicated cataract surgery) were excluded. Preoperative data included visual acuity and donor endothelial cell density by specular microscopy. The main outcome measures were postoperative best spectacle-corrected visual acuity (BSCVA) and endothelial cell loss at 3, 6, and 12 months.


There were 59 eyes with Fuchs dystrophy and 41 eyes with PBK. In eyes without vision-limiting pathology, the median postoperative BSCVA was consistently 20/40 (range 20/20-20/400) at 3 months (n = 61 eyes), 6 months (n = 55 eyes), and 12 months (n = 48 eyes). Endothelial cell loss was 13.7% at 3 months (n = 57), 13.5% at 6 months (n = 61), and 14.9% at 12 months (n = 53). Primary graft failure occurred in 1 eye, attributable to incorrect use of the insertion device. Two eyes with complete donor dislocation were rebubbled successfully. The most common complication was glaucoma/ocular hypertension in 29 eyes (34.1%) without prior glaucoma and treatment escalation in 6 eyes (40.0%) with prior glaucoma. Of the 78 eyes with 12 months follow-up, 2 (2.6%) developed endothelial rejection, and 1 (1.3%) subsequently failed.


The use of this donor insertion device during DSAEK demonstrates good outcomes and potentially low endothelial cell loss at up to 12 months after surgery.

In recent years, Descemet stripping automated endothelial keratoplasty (DSAEK) has become a proven alternative to penetrating keratoplasty (PK) in the treatment of endothelial dysfunction. The advantages of DSAEK include a faster postoperative recovery and visual rehabilitation, minimal astigmatic change, and retention of the cornea’s tectonic strength. However, DSAEK is also associated with a higher rate of iatrogenic primary graft failure, as well as a higher initial drop in the endothelial cell density (ECD) when compared to PK. These may be attributed to the increased donor handling and manipulation required during surgery, especially if the donor is folded and inserted with forceps. While the rate of endothelial cell loss appears to stabilize much earlier when compared to PK, long-term graft survival in DSAEK remains to be determined.

In order to minimize graft manipulation, various donor insertion techniques have been developed. This includes the use of devices such as the EndoGlide (AngioTech, Reading, Pennsylvania, USA/Network Medical Products, North Yorkshire, UK), an FDA-approved device for donor insertion during DSAEK. We described the early clinical results of DSAEK with this donor insertion device in a small series of Asian eyes and found a low rate of postoperative complications and a low endothelial cell loss at 6 and 12 months.

This paper presents the results of DSAEK with this donor inserter in 100 eyes with either Fuchs endothelial dystrophy or pseudophakic bullous keratopathy (PBK), without any other anterior segment pathology or previous intraocular surgery except for uncomplicated cataract surgery with lens implantation. The study was designed to establish the rate of surgical complications and postoperative endothelial cell counts after DSAEK with the EndoGlide, in the absence of other factors that could adversely affect endothelial cell function or graft survival.


Our study was a retrospective interventional case series of 100 consecutive eyes with visually significant corneal edema from Fuchs endothelial dystrophy or PBK that underwent DSAEK with the use of the EndoGlide donor insertion device at a single study center, the Singapore National Eye Centre (SNEC). Data for this study were obtained from the Singapore Corneal Transplant Study, which is an ongoing prospective cohort study tracking all corneal transplants performed at the SNEC since 1991. Both the Singapore Corneal Transplant Study and this present study were conducted with the approval of the Institutional Review Board of the Singapore Eye Research Institute and conformed to the tenets of the Declaration of Helsinki.

For this study, we excluded eyes with the presence of any anterior segment pathology (eg, anterior chamber intraocular lenses [IOLs], primary angle-closure glaucoma) and almost all previous intraocular surgery (eg, trabeculectomy, glaucoma drainage devices, failed corneal grafts), as well as cataract surgery with complications (eg, capsular rupture, placement of a sulcus IOL); however, corneal edema from previous uncomplicated cataract surgery with implantation of a posterior chamber IOL was included in our study.

Prospective data collected before surgery included demographic details and prior ophthalmic history. Ethnicity was classified into the following 4 categories, as defined by the Singapore Department of Statistics : Chinese, Malay, Indian, and Others. Intraoperative and postoperative complications, including iatrogenic primary graft failure, graft dislocation, endothelial rejection, and graft failure, were documented. Outcome measures were best spectacle-corrected visual acuity (BSCVA) and central ECD before surgery and at 3 months, 6 months, and 12 months after surgery. Intraocular pressure (IOP) was also measured at these postoperative visits. Visual acuity was assessed under standardized conditions and documented in Snellen notation using established guidelines.

Preoperative donor ECD for all tissues was based on specular microscopy performed by certified technicians at the Singapore Eye Bank with the Konan Keratoanalyzer EKA-98 (Konan Medical Corp, Hyogo, Japan). Postoperative ECD was performed using the Noncon Robo Specular Microscope NSP-9900 (Konan Medical Corp) in the SNEC by independent ophthalmic technicians trained in specular microscopy. Calibrations and magnifications were standardized and performed as previously described. The ECD was considered acceptable if at least 50 endothelial cells were counted and marked on a high-quality image of the central corneal endothelium; otherwise, the ECD was excluded from the results. Postoperative endothelial cell loss was calculated as a percentage of the preoperative donor ECD.

Statistical Analysis

Descriptive statistics for normally distributed variables are reported as mean ± standard deviation; otherwise, median and range are reported. The 95% confidence intervals (95% CI) are reported along with the mean 3-month, 6-month, and 12-month percentage ECD loss. Comparisons between categorical variables were conducted by χ 2 tests, and 1-way analysis of variance was used for means. Statistical significance was set at P < .05. All data analysis was carried out with SPSS Statistics 17.0 (SPSS Inc, Chicago, Illinois, USA).

Surgical Technique

The surgical technique used in this study has previously been described in detail. In brief, a 4.5-mm scleral tunnel incision was created temporally, and phacoemulsification and lens implantation with a cohesive viscoelastic was performed at this stage, if required. The Descemet membrane was then stripped with the cohesive viscoelastic in the anterior chamber (AC), or under air. An AC maintainer was preplaced, and an inferior peripheral iridotomy and 4 venting incisions were made at this juncture. The prepared donor tissue was then loaded into the device and inverted 180 degrees prior to insertion, so that the endothelial surface faces down. The device was inserted through the sclera incision and the graft pulled into the AC with forceps introduced through a nasal paracentesis. Once the graft was fully uncoiled, a full air tamponade was maintained for at least 6 minutes.

Postoperative Care

After surgery, all patients received a standard postoperative regime of a topical steroid, either prednisolone acetate 1% (Pred Forte; Allergan, Inc, Irvine, California, USA) or dexamethasone sodium phosphate 0.1% (Minims; Bausch & Lomb, Inc, Rochester, New York, USA), and a topical antibiotic. The steroid was applied every 3 hours for a month, then 4 times daily for 2 months, and then tapered by 1 drop every 3 months down to 1 drop per day by the first year. The steroid was continued indefinitely thereafter. In eyes with a presumed steroid-related increase in IOP to more than 21 mm Hg, surgeons had the discretion to taper the frequency of the topical steroid more rapidly or to switch to a less potent steroid, such as loteprednol etabonate (0.5%) (Lotemax; Bausch & Lomb, Inc).


Of the 197 eyes that had DSAEK performed at the SNEC between January 1, 2009 and June 30, 2011, we included 100 eyes of 95 patients that fulfilled the study criteria. There were 40 male patients, the mean age was 69.4 (SD ± 9.2) years old, and there were 66 Chinese patients, 8 Malay patients, 3 Indian patients, and 18 patients of other ethnicities.

There were 52 phakic eyes with Fuchs endothelial dystrophy; all had some degree of cataract and so had combined phacoemulsification and intraocular lens implantation at the time of DSAEK surgery. In 41 eyes with PBK and 7 pseudophakic eyes with Fuchs endothelial dystrophy and corneal edema, DSAEK alone was performed. Patients with Fuchs dystrophy were significantly younger than those with PBK (mean age 66.7 years old vs 73.4 years old, P < .05). Of the 100 grafts transplanted, 63 were in the right eye, preoperative donor ECD averaged 2806 ± 218 cells/mm 2 , and median graft diameter was 8.75 mm (range 8.0-9.5 mm). The average graft thickness was 177 ± 43 μm (measurements were available for 92 grafts). The surgeries were performed by 5 attending corneal surgeons, all with previous experience performing DSAEK with the sheets glide method for donor insertion.

Of the 100 eyes included in the study, 96 eyes completed 3 months of follow-up, 94 eyes completed the 6-month follow-up, and 78 eyes had follow-up data at 12 months. A comparison between the eyes that completed 12 months of follow-up vs the eyes that did not found no statistically significant differences in baseline parameters such as age ( P = .305), sex ( P = .189), ethnicity ( P = .754), preoperative diagnosis ( P = .192), graft diameter ( P = .205), or preoperative donor ECD ( P = .754).

Intraoperative Complications

All cases of DSAEK had successful donor insertion with this device and without need for alternate donor insertion techniques (eg, the sheets glide method). However, donor preparation was eventful in 3 eyes. In the first case, the donor tissue was caught in the blade of the Hanna trephine during trephination and displaced off the punch block. As there was no standby tissue available, the graft was inserted and the DSAEK completed uneventfully. Fortunately, the donor appeared not to have suffered significant endothelial trauma, and remained clear at 1 year with an ECD loss of 14.7%. In a second case, the graft was accidentally flipped over on the punch block after trephination and standby tissue was used to complete the surgery successfully. In the third case, there was an uneven lamellar cut with the automated lamellar therapeutic keratoplasty system (Moria, Antony, France) used during donor cornea preparation, and manual dissection was required to achieve a uniform thickness of 200 μm. The rest of the surgery was uncomplicated. No difficulties or surgical events were encountered during the donor coiling or insertion stages with the donor inserter, with the exception of 1 case of incorrect device insertion, which was not noted at the time of surgery (see below).

In 2 eyes with Fuchs endothelial dystrophy and cataract, posterior capsular rupture occurred during phacoemulsification of the cataract, and both required anterior vitrectomy and the placement of the IOL in the sulcus. Subsequent donor insertion and completion of the DSAEK was performed without incident.

Early Postoperative Complications

In the immediate postoperative period, 1 primary graft failure was noted. Review of the surgical video revealed that this was the result of incorrect device insertion: the device with the loaded donor tissue was not inverted prior to insertion into the AC and thus the graft was pulled into the recipient eye with the endothelium facing the corneal stroma. This eye underwent a successful graft exchange with the donor insertion device 10 days later, and at 1 year the graft remained clear, with an endothelial cell loss of 12.5%.

Two eyes with complete donor dislocation were rebubbled successfully within the first week of surgery. In 1 eye, dislocation was attributable to wound leak from the main incision with resulting postoperative hypotony, while in the second eye no clear reason could be identified. Another 4 eyes were noted to have small, peripheral detachments after surgery and were observed; all reattached without intervention.

Visual Acuity Results

In the analysis of the postoperative results, we excluded the eye with iatrogenic primary graft failure, and note that 3 patients (3 eyes) defaulted before completing 3 months of follow-up. The median preoperative BSCVA was 20/80 (range 20/25-20/20 000). A total of 96 eyes completed 3 months of follow-up, and the median BSCVA was 20/50 (range 20/25-20/2000). Ninety-four eyes completed the 6-month follow-up, and the median BSCVA was 20/40 (range 20/20-20/20 000). At 12 months, 78 eyes had completed follow-up, and the median BSCVA was stable at 20/40 (range 20/20-20/20 000).

After eyes with vision-limiting pathology (eg, macular or retinal disease, optic neuropathy) were excluded from the visual acuity analysis, the median BSCVA before surgery was 20/70 (range 20/25-20/20 000) in 63 eyes. After DSAEK, the median BSCVA was consistently 20/40 at 3 months (n = 61 eyes), 6 months (n = 55 eyes), and 12 months (n = 48 eyes); the BCSVA ranged from 20/20-20/400 at these time points after surgery. In terms of percentage of eyes achieving BCSVA of 20/40 or better, the result was 22.2% (14 of 63 eyes) before DSAEK, 60.7% (37 of 61 eyes) at 3 months, 71.0% (39 of 55 eyes) at 6 months, and 85.4% (41 of 48 eyes) at 12 months.

Endothelial Cell Density

Out of 57 eyes with a recordable central ECD at 3 months, the mean ECD was 2450 ± 519 cells/mm 2 and the mean endothelial cell loss was 13.7% (95% CI, 9.15%-18.20%). At 6 months, there were 61 eyes with a recordable ECD reading; the mean ECD was 2421 ± 494 cells/mm 2 and the mean cell loss was 13.5% (95% CI, 9.40%-17.63%). At 12 months, 53 eyes had a recordable ECD reading, with a mean ECD of 2365 ± 463 cells/mm 2 and a mean endothelial cell loss of 14.9% (95% CI, 10.52%-19.35%).

When we compared eyes with Fuchs endothelial dystrophy vs PBK, we found no significant difference in the mean preoperative donor ECD (2814 cells/mm 2 vs 2778 cells/mm 2 , P = .422) and no significant difference in the postoperative endothelial cell loss at 3 months (13.9% vs 13.4%, P = .913), 6 months (14.0% vs 12.7%, P = .761), and 12 months (14.1% vs 16.3%, P = .628).

The majority of the cases (78 eyes) were performed by the 2 senior authors (J.M. and D.T.) and the remaining cases by 3 other attending corneal surgeons who were still in the learning curve of using the EndoGlide. A comparison of the postoperative cell loss between these 2 groups of surgeons (senior surgeons vs the 3 other surgeons) found no statistically significant differences at 6 months (12.9% vs 19.1%, P = .38) and 12 months (15.6% vs 9.4%, P = .37). However, the number of eyes with analyzable cell counts contributed by the 3 other surgeons amounted to only 6 eyes at 6 months and 12 months, compared to the more than 46 eyes by the 2 senior authors.

Late Postoperative Complications

Glaucoma/ocular hypertension

Preoperatively, there were 15 eyes with a pre-existing diagnosis of glaucoma or ocular hypertension and requiring at least 1 glaucoma drop for IOP control. After DSAEK, 1 eye with ocular hypertension no longer required medical treatment, 4 of these 15 eyes (26.7%) required an increase in the number of drops to achieve the target pressure, and 2 other eyes (13.3%) required trabeculectomy with mitomycin C for uncontrolled pressures despite maximal medical therapy.

Out of 85 eyes without prior glaucoma or ocular hypertension, 29 (34.1%) developed an IOP rise above 21 mm Hg after surgery, and required at least 1 glaucoma drop for IOP control despite tapering the frequency and/or changing the potency of the topical steroids. All have been medically controlled and none have required glaucoma surgery thus far.

Corneal complications

One patient with PBK had visually significant central folds in an otherwise thin and functional graft after uncomplicated DSAEK. Despite a best-corrected visual acuity of only 20/400, the patient declined a graft exchange to improve the vision.

Graft rejection and late graft failure

Excluding our 1 eye with iatrogenic primary graft failure, there were 21 eyes that did not complete 12 months of follow-up. None had evidence of endothelial rejection and all had clear, functioning grafts at the last documented review.

Of the 78 eyes that have completed 12 months of follow-up, there have been 2 eyes with endothelial rejection, occurring at 3 months and 6 months, respectively. In the first eye, the rejection episode resolved rapidly with intensive topical steroids and the cornea remained clear at 12 months. In the second eye, the patient was treated with intensive topical prednisolone acetate 1%, and the cornea remained clear until the 12-month review, when inferior corneal edema was noted. This patient declined further surgery despite the impending graft failure and a drop in the vision to 20/400. Apart from this 1 eye, no other eye has developed late endothelial failure thus far.

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Jan 9, 2017 | Posted by in OPHTHALMOLOGY | Comments Off on Descemet Stripping Automated Endothelial Keratoplasty With a Donor Insertion Device: Clinical Results and Complications in 100 Eyes

Full access? Get Clinical Tree

Get Clinical Tree app for offline access