Purpose
To determine the risk factors for Descemet Stripping Automated Endothelial Keratoplasty (DSAEK) graft detachment and graft failure to refine and standardise surgical methods.
Methods
673 patients who underwent DSAEK in Moorfields Eye Hospital since 2010 were included in this study. In this series, all endothelial grafts were surgeon-cut. Clinical data on the following were collected and analysed: patient demographics, indication for surgery, pre-operative co-morbidities, best-corrected visual acuity (BCVA), refractive status, grade of the primary surgeon, donor material characteristics, method of tissue cutting, insertion technique, the presence of stromal vents, duration of air tamponade, graft detachment rate and graft failure rate.
Results
The commonest indications for DSAEK were pseudophakic bullous keratopathy (39.9%) and Fuchs’ endothelial dystrophy (35.4%). 12.7% were re-do DSAEK procedures, 25.9% were combined with cataract surgery and the remainder were isolated DSAEK operations. 73.4% had prior ophthalmic surgery. Mean BCVA improved from 0.85±0.61 LogMAR to 0.50±0.60 LogMAR. The graft detachment rate was 19% with a refloat success rate of 89.6%. Graft failure rate at 1 year was 8.9%, rising to 15.8% at year 2. Univariate regression analysis failed to identify any significant risk factors for graft detachment. Previous glaucoma surgery, lens-iris diaphragm disruption, donor age, a surgeon caseload of ≤20 DSAEK procedures per year, the suture pull through method of graft insertion and graft detachment were all associated with graft failure. Multivariate regression modeling showed that previous glaucoma surgery, lens-iris diaphragm disruption and graft detachment are independent predictors of graft failure.
Conclusions
Previous glaucoma surgery and lens-iris diaphragm disruption are risk factors associated with endothelial graft failure. These factors should be considered during patient counseling before offering DSAEK surgery for endothelial failure.