The most common reported complication following DSAEK surgery is graft dislocation (Fig. 12.3), which has been reported in 1.5–35 % of cases [17, 20, 35, 39, 40, 42, 43]. The causes for graft detachments and complete dislocations vary widely, and surgeons have published mixed results regarding the association between previous glaucoma surgery and DSAEK graft dislocation. Some surgeons report no significant difference in dislocation (15 % in eyes with a history of glaucoma surgery vs 13 % without – Wiaux et al. 2011), while others report a significant association (9 % dislocation with a history of glaucoma surgery vs 2 % with no history of glaucoma surgery – Goshe et al. [10]). It is likely that both patient demographics and surgical technique play a role. However, within the group of patients with a history of glaucoma surgery, there appears to be a strong correlation between postoperative hypotony and graft detachment [10]. The surgical learning curve and associated trouble with graft manipulation and air bubble management have been shown to lead to higher detachment rates in early cases which may decrease over time [30]. Other risk factors to which graft detachment is often attributed include eye rubbing, interface fluid remaining at the end of surgery, retained interface viscoelastic, and folds of retained Descemet membrane. DSAEK surgery for penetrating keratoplasty failure presents the unique challenge of matching the DSAEK lenticule to the posterior graft–host junction of the full-thickness corneal graft (Fig. 12.4). Sizing the graft so that the diameter is smaller than the penetrating graft and careful centration during surgery has been associated with a low rate of postoperative graft dislocation [38] However, Price et al. have also shown that making the graft larger than the diameter of the overlying PK has also yielded a low detachment rate [28]. The key to success appears to be to not make the DSAEK graft the same diameter as the overlying failed PK in order to avoid the edge protuberances of the PK edges from pushing the graft edges down and creating edge separations.

The management of partial or complete graft detachment is the placement of a new air bubble in the anterior chamber with subsequent face-up patient positioning to support the graft tissue. This is performed either in a minor procedure room or in the operating room. The operating room may be preferred in the case of the need for more significant surgical manipulation. Graft reattachment occurs in the majority of grafts which undergo rebubble [30, 39, 42, 43]. In the minority of eyes which do not achieve successful attachment after several rebubble attempts, iatrogenic primary graft failure is often declared, and either a repeat DSAEK or PKP can be performed.

Failure of initial corneal clearance after any form of endothelial keratoplasty is usually attributable to excessive manipulation of the donor tissue during surgery, and this has been termed iatrogenic primary graft failure. If donor characteristics or tissue preparation techniques are suspected of being responsible for graft failure, then true primary graft failure can be declared in the absence of excessive surgical manipulation. The rates of iatrogenic primary graft failure vary widely in the published DSAEK reports from as low as 0 % [42, 43] to as high as 17 % [39, 40]. In most cases graft replacement with a new DSAEK lenticule is the treatment of choice. However, in situations where patient comorbidities may be responsible for increased risk of dislocation such as the inability to maintain anterior chamber air due to glaucoma surgery or a history of vitrectomy in a unicameral eye, the procedure of choice for DSAEK graft failure may be penetrating keratoplasty.

The placement of a large air bubble in the anterior chamber at the end of the DSAEK surgery may lead to pupil block and occasionally permanent vision loss due to optic neuropathy or a fixed dilated pupil due to Urrets-Zavalia syndrome [5]. Although most large series have not reported pupil block after DSAEK, there have been some reports of this complication in the literature [39, 40]. The routine use of mydriatic drops at the end of the surgery and titration of an air bubble that can clear the edge of the dilated pupil has been associated with elimination of this complication in a large series [42, 43]. The placement of an inferior peripheral iridectomy prior to placement of the air bubble may also prevent pupil block. Cases of pupil block may be treated with the release of air from the anterior chamber. However, it is important to identify the complication early in its course to prevent prolonged elevated intraocular pressure and its sequelae.

Endophthalmitis following DSAEK surgery is exceedingly rare, but has been reported (Fig. 12.5). Responsible organisms include Mycobacterium abscessus [7], Candida spp. [8], and Streptococcus pneumoniae [35]. A unique characteristic of postoperative infections following DSAEK is the possibility of keratitis involving the graft–host interface. Possible treatments include topical antimicrobials, injection of antimicrobials into the interface, or removal of the DSAEK lenticule with placement of a new DSAEK graft or penetrating keratoplasty. Theoretically, donor rim cultures could be used to guide therapy of post-DSAEK infections; however, this topic remains controversial.

Fungal infections following a DSAEK transplant are exceedingly rare, but have been reported [19]. These infections tend to involve the graft–host stromal interface [24]. The most commonly reported causative organism is Candida albicans. Depending on local climate, the risk may increase with increased exposure by donors. Many surgeons will test donor tissue for the presence of potential pathogenic microbes by culturing the donor corneal-scleral rim tissue left over after trephination, but often the culture reports will only be available a week or more after transplantation. Even in cases where a positive fungal growth from the donor rim tissue is identified, an infection usually will not develop. However, in the face of a positive donor rim culture and a clinical picture of increasing inflammation of the eye, early identification of the causative organism allows faster and more appropriately targeted therapy. If a fungal infection is identified by culture and clinically manifests in the interface or graft itself, eradication may require a combination of topical, oral, and intracameral antifungal therapy and often ultimately results in removal of the infected graft.

Endothelial cell density decreases over time following DSAEK surgery, with the greatest decline occurring in the first year after surgery. Continued decline at a slower rate has been demonstrated up to 5 years following transplantation. Late graft failure has been reported in 6–8 % of patients 5 years after undergoing DSAEK [28, 32]. At 5 years after surgery, patients with a history of pseudophakic or aphakic bullous keratopathy have a greater risk of late graft failure (24 %) compared to patients with Fuchs dystrophy (5 %). In addition, long-term graft survival in eyes with prior glaucoma surgery is worse than with Fuchs dystrophy eyes [28]. Contrary to surgeon biases, the characteristics of donor tissue that passes EBAA standards have no relationship to graft survival or postoperative cell counts. Tissue that is “young, fresh, and with a high cell count” does not improve the postoperative outcomes or health of the DSAEK graft. In large studies, preoperative donor endothelial cell density, death-to-preservation time, donor age, and recipient age have not been associated with decreased postoperative endothelial cell density or late graft failure [42, 43]. The most important preventable causes of late graft failure are endothelial cell loss from traumatic surgical implantation and immunologic rejection. Late graft failures may be reduced by adhering to careful surgical technique with a 5-mm incision, gentle treatment of the endothelial layer, and close postoperative care to monitor for and treat immunologic rejection episodes as expediently as possible.

Endothelial immunologic rejection is an important treatable cause of endothelial cell loss and late graft failure. Cumulative rates of rejection have been reported at 1 year between 6.0 and 7.6 %, at 2 years between 12 and 14 %, and at 3 years 22 % [15, 29, 48]. The peak time for immunologic rejection occurs between 12 and 18 months following DSAEK surgery [21]. Patients generally present with corneal edema, keratic precipitates, and decreased visual acuity (Fig. 12.6). Approximately one third of patients are asymptomatic at the time of presentation. Risk factors for immunologic rejection include stopping topical corticosteroids and being of African-American descent [29]. Endothelial rejection is treated with frequent topical corticosteroids and can be followed with pachymetry and clinical exam. Most patients are treated successfully with medication alone, but 7–26 % eventually require graft replacement. Some studies have suggested that DSAEK grafts are less likely to fail as a result of immunologic rejection than PKP grafts [4]. In order to minimize the occurrence of immunologic rejection following DSAEK surgery, we recommend continuing corticosteroid therapy until at least 2 years following transplant, strict return precautions for telltale symptoms of rejection, and monitoring patients frequently during the first 2 years after surgery to ensure that they do not suffer significant cell loss from subclinical rejection episodes.