Making the transition from PK to DSEK: experiences during the learning curve


Eydís Ólafsdóttir, MD
Department of Ophthalmology
University of Iceland, National University Hospital
101 Reykjavík, Iceland
Tel: + 354 543 1000
Fax: + 354 543 4831


Purpose:  To report our experiences during the transition from penetrating keratoplasty to Descemet’s stripping with endothelial keratoplasty (DSEK).

Methods:  All patients undergoing DSEK during the period of April 2008 to April 2009 were included in this study.

Results:  All grafts were attached and clear at both the 6- and 12 -month follow-up examinations. Mean best-corrected visual acuity was 0.6 at 6 months and 0.7 at 12 months for patients without other ocular comorbidity.

Conclusion:  With adequate attention to detail, DSEK seems to be a safe and successful treatment for corneal endothelial disease, also during the surgeon’s learning curve.


An alternative to penetrating keratoplasty (PK) for corneal endothelial disease has been sought by corneal surgeons for many years. Several different techniques have been tried, with varying levels of surgical complexity and results (Tillett 1956; Melles et al. 1998, 1999; Ehlers et al. 2000; Azar et al. 2001; Terry & Ousley 2001).

Descemet’s stripping with endothelial keratoplasty (DSEK), introduced by Gerrit Melles (Melles et al. 2004) and named and popularized in the US by Francis Price (Price & Price 2005, 2006), has during recent years become the method of choice for endothelial replacement. Several studies have shown the safety and efficacy of the technique (Lee et al. 2009). Despite its growing popularity worldwide, there is a paucity of peer-reviewed studies on DSEK from the Nordic countries, except from Denmark (Hjortdal & Ehlers 2009), as well as descriptions of what to expect while making the transition from PK.

Our department began doing DSEK in April 2008, and I will report our experiences during the learning curve of this procedure.

Methods and Material

All patients undergoing DSEK during the period of April 2008 to April 2009 were included in this study (Table 1). All procedures were performed by the same surgeon (EÓ). The patients were examined by the surgeon preoperatively, as well as 6 and 12 months after surgery. Best-corrected visual acuity (BCVA) measured with a Snellen chart was recorded, as well as a description of the clarity and location of the graft. Ocular comorbidity and postoperative complications were also recorded.

Table 1.   Characteristics of the study participants.
6 m
1 year
ComplicationsClear graft
  1. BCVA, Best-corrected visual acuity.

82MBullous keratopathyAMDHM<0.1<0.1+
53MBullous keratopathyCongenital glaucoma/ShuntHM0.10.1+
75FFuchsSubepithelial fibrosis0.20.20.4+

Eleven eyes of 11 patients, five men and six women, were operated during this time period. Mean age at surgery was 78 years, with a range of 53–84 years. The indication for the procedure was Fuchs’ endothelial dystrophy in nine patients and bullous keratopathy in the other two. All eyes were pseudophakic at the time of the DSEK operation or were made pseudophakic by having cataract surgery 1 month prior to undergoing DSEK.

Five patients showed no other ocular comorbidity (all of them with Fuchs’). Four patients had age-related macular degeneration in both eyes. One patient with Fuchs’ had subepithelial fibrosis. One of the patients with bullous keratopathy also suffered from congenital glaucoma and had received an Ahmed shunt previously. The other eye in this patient was amaurotic.

All the grafts originated from the North Carolina Eye Bank and were precut with microkeratome by eye bank technicians 2–3 days prior to transplantation.

In the first two surgeries, forceps were used to introduce the graft into the anterior chamber; the Busin glide was used in all subsequent procedures. In the majority of cases, surgical technique involved a 5 -mm incision, anterior chamber maintainer, Descemet’s stripping, forceps pull-through insertion of the graft assisted by the Busin glide, injection of an air bubble, four venting incisions and positioning of the graft with a LASIK-roller. A complete air fill of the anterior chamber was maintained for 20 min, then part of the air was removed and replaced with balanced salt solution.

Meticulous attention to details of the surgical steps was observed during each of the procedures. No intraoperative complications, affecting the outcome, occured.


All grafts were attached and clear at both the 6- and 12 -month follow-up examinations. Mean best-corrected visual acuity was 0.6 at 6 months and 0.7 at 12 months for patients without other ocular comorbidity. BCVA in all patients improved after surgery. The patient with Fuchs’ dystrophy and preoperative subepithelial fibrosis limiting postoperative BCVA, later underwent photorefractive keratectomy (PRK). Visual acuity 1 month after PRK was 0.4 in this patient.

One patient needed rebubbling, which was performed on the third postoperative day. One case of graft rejection occured, in the same patient that needed rebubbling. The rejection occured 1 year after primary surgery and 1 month after topical steroids had been discontinued. The rejection episode was successfully treated with topical steroids and the graft remains clear. There were no cases of primary endothelial failure or iatrogenic glaucoma.


Descemet’s stripping with endothelial keratoplasty has several advantages over PK – including early visual recovery, better refractive stability and postoperative refractive outcomes, less wound and suture-related complications and less intraoperative and late choroidal haemorrhage risk (Lee et al. 2009; Price & Price 2010a). Several studies have been published describing the visual results of surgery for Fuchs’ dystrophy and bullous keratopathy. BCVA of 0.5 or better was achieved with endothelial keratoplasty in 81%, 69% and 77% of patients with Fuchs’ or bullous keratopathy, when compared to 54%, 64% and 65% with PK in patients with Fuchs’ (Price et al. 1991; Pineros et al. 1996; Claesson et al. 2002; Gorovoy 2006; Price & Price 2006; Chen et al. 2008).

However, the DSEK procedure itself is more complex, possibly affecting the results during the surgeon’s learning curve (Gorovoy 2006; Price & Price 2006; Koenig et al. 2007; O’Brien et al. 2008).

The key elements of successful DSEK surgery are inserting the donor tissue without damaging the endothelium, getting it to adhere, and to achieve long-term survival of the graft. This can be difficult for the novice surgeon, but choosing adequate surgical techniques and tools are of great help. Forceps insertion threatens to cause more intraoperative trauma to the endothelium of the donor disc, which is why I converted to the Busin glide after my first surgeries (Busin et al. 2008; Bahar et al. 2009). The use of an anterior chamber maintainer will assist in keeping the anterior chamber deep, making an atraumatic insertion possible. An additional benefit of this device is that its fluid stream helps the graft to unfold once inside the eye. The edge of the donor disc needs to be grasped with the pull-through forceps until the graft has unfolded and the anterior chamber infusion has been turned off. Failing to observe this step might cause the graft to prolapse out of the wound (Price & Price 2010a).

For adherence and clarity, it is important that no fluid or foreign material is present in the host–donor interface. In one of my cases, blood was sequestered in the interface during surgery, but could be flushed out with BSS injected through the small incision made for scoring the host endothelium, followed by attachment of the disc through injection of an air bubble. Placing the scoring-incision in a position where it is covered by the graft increases its safety and versatility.

For special cases, like my patient with a glaucoma shunt, achieving attachment can be more difficult. Taking the time needed to obtain a good air fill is mandatory, the air can then be left in place as has been described in the literature (Price & Price 2010a).

Apart from the risk of intraoperative technical difficulties, concerns have also been voiced over the rate of endothelial cell loss during this procedure, and of the long-term survival of DSEK grafts. Recent studies by Price et al. (Price & Price 2010b) have shown less endothelial cell loss after 5 years in DSEK than PK, and a similar rate of 5-year graft survival for the two methods. Shamie et al. (2010) have shown that visual acuity after DSEK continues to improve up to 3 years after surgery.

The most common complications to DSEK are graft dislocation, endothelial graft rejection, primary graft failure and iatrogenic glaucoma. According to the literature (Lee et al. 2009), graft dislocation occurs on average in 15% of cases, graft rejection in 10%, primary graft failure in 5% and iatrogenic glaucoma in 3%.

Results and complication rates of this small study are similar to what has been observed in larger materials presented by experienced surgeons.

In conclusion, with adequate attention to detail, DSEK seems to be a safe and successful treatment for corneal endothelial disease, also during the surgeon’s learning curve.