6 Ultrathin Descemet Stripping Automated Endothelial Keratoplasty Endothelial keratoplasty (EK) has revolutionized the field of corneal transplantation for focal endothelial disease by providing a safer procedure with more predictable refractive outcomes and less risk for rejection compared to penetrating keratoplasty.1 Since Tillett’s first discussion of posterior lamellar keratoplasty in 19562 and Melles’s article on endothelial keratoplasty in 1998,3 several variations have been proposed to improve refractive outcomes and minimize rejection. These include deep lamellar endothelial keratoplasty (DLEK),4 Descemet stripping endothelial keratoplasty (DSEK),5 Descemet stripping automated endothelial keratoplasty (DSAEK),6 Descemet membrane endothelial keratoplasty (DMEK),7 and ultrathin Descemet stripping automated endothelial keratoplasty (ultrathin69 DSAEK)8. The postoperative outcomes after standard DSAEK typically result in visual acuities better than 20/409,10; however, in some instances, despite a clear graft with no apparent irregularities, the final visual acuity is worse than expected. Two main factors have been proposed to explain this variability in results: graft thickness11,12,13 (preoperative and postoperative) and the higher-order aberrations (HOAs) induced by irregularities in the donor–host interface and in tissue architecture.6,14,15,16 Aside from a thinner lenticule, the double-pass microkeratome technique used for ultrathin DSAEK seems to provide better tissue architecture than a single-pass preparation, with comparable postoperative outcomes to DMEK and without the steep learning curve that DMEK requires to achieve proficiency.17,18 In ultrathin DSAEK the lenticule is also significantly thinner than DSAEK, but the surgical technique remains essentially the same. Despite excellent postoperative outcomes with DMEK19 there is still poor traction for this procedure due to the significant learning curve and technical challenges DMEK presents. The comparative efficacy of DMEK compared to ultrathin DSAEK has yet to be fully elucidated but may be more similar than outcomes for regular-thickness DSEK grafts. In 2011 Neff et al published their findings of thinner lenticules, especially those < 131 µm, achieving better visual acuities than thicker grafts (20/25 in 100% and 20/20 in 71%).20 There are many reports disputing the association between preoperative lenticule thickness and postoperative outcomes.21,22,23,24,25 In these reports, preoperative and postoperative donor thicknesses were > 130 µm (> 145 µm in preoperative and >141 µm in postoperative measurements); and when preoperative graft thicknesses < 125 µm were analyzed separately, the percentage of patients with better visual acuities trended toward better visual outcomes.21,24,26,27 In several publications where only thinner lenticules have been studied the results are still conflicting. Some show a significant improvement in visual outcomes and rate of detachment and dislocation when lenticules are < 130 µm.28,29 Others have shown no significant correlation between postoperative graft lenticule thickness and visual outcomes.30,31 The current evidence associating graft thickness with postoperative visual outcomes is inconclusive, and current available studies do not yet present robust data and statistical analyses to prove or disprove this relationship,32 specifically with lenticules < 130 µm. There is an average hyperopic shift of + 1.0 D after DSAEK.33,34,35 This seems to be related to tissue architecture and the negative lens effect that lenticules have, which is due to increased peripheral lenticule thickness.36 In correlation analyses, a thicker lenticule center, a larger lenticule diameter, and negative curvature profiles have all been correlated with a higher refractive shift,27,34 with stronger correlation when a thicker lenticule and a negative lenticule curvature profile are combined.27,36 Despite negligible changes in the anterior corneal topography after all types of endothelial keratoplasty,15,27,36 there is an increase in posterior corneal HOAs after both DSAEK13,15,29,35,37 and, to a lower degree, DMEK.38 No studies are available in regard to HOAs after ultrathin DSAEK. In 2006 Melles et al introduced DMEK,7 limiting the transplanted lenticule to Descemet membrane and endothelium only. DMEK likely represents the thinnest possible tissue routinely amenable to transplantation. However, DMEK poses many technical challenges and a significant learning curve to the surgeon.17,18 There are some instances in which DMEK lenticule unfolding, which requires specific aqueous dynamics, would be challenging if at all possible, due to complexities from comorbidities or preexisting surgeries. In these patients DSAEK remains the procedure of choice, and a thinner lenticule may be desirable to maximize the postoperative outcomes. In 2012 Busin et al8 published their concept of ultrathin DSAEK previously presented at the 2009 European Society of Cataract and Refractive Surgery meeting and proposed the use of a preoperative lenticule thinner than 100 µm by using the “microkeratome double-pass technique” for donor preparation, which would standardize the preparation of grafts, resulting in consistently thinner and more planar lenticules.39,40 Ultrathin DSAEK is similar to standard DSAEK in terms of the surgical procedure, with the difference being mainly in donor preparation for a lenticule ideally < 100 µm. Even though ultrathin lenticules can still be obtained with a single-pass micro-keratore by applying a nomogram for microkeratome head selection,41,42 the double-pass microkeratome technique yields a more planar surface that offers some advantages in postoperative quality of vision and refractive shift. Ultrathin DSAEK offers the advantages of minimizing stroma attached to Descemet membrane and endothelium in the donor lenticule while maintaining essentially the same surgical technique that has been used for standard DSAEK, with lenticule thickness measurements as low as 23 µm by 6 months postoperatively.42 The concerns for implementation of ultrathin DSAEK lie in tissue manipulation and increased tissue flexibility, which may result in postoperative irregularities in both the stroma interface and the endothelial side that may induce HOAs21; however, these concerns could also be applied to the very thin lenticule used in DMEK. Studies looking at the relationship between DSAEK graft folds and HOAs have shown that thicker grafts actually tend to have more folds and induce more HOAs43; however, reports on HOAs after ultrathin DSAEK are not available. The main objectives in the donor preparation for endothelial keratoplasty are to achieve a specific target lenticule thickness and to obtain the most planar cut possible in order to minimize postoperative refractive changes while inflicting minimal trauma to the endothelial cells in the graft. Ultrathin DSAEK lenticules can be obtained both with a single-pass and with a doublepass microkeratome technique. Different strategies have been suggested in order to consistently obtain thinner cuts with a single microkeratome pass, such as stromal dehydration,44 development of a nomogram for microkeratome head setting based on central corneal thickness prior to the cut,42 and adjustment of the translational speed of the microkeratome in thicker donors, since slower translational speeds result in deeper cuts and thinner lenticules.40 With a single microkeratome cut there is higher variability in thicknesses throughout the lenticule compared to two microkeratome passes.26,39 This variability results in lenticules that are thinner in the center than in the periphery,40 which has been related to the postoperative hyperopic shift observed after DSAEK.36 Additionally, the predictability of the final lenticule thickness decreases as the target thickness decreases, dropping from 78% of cuts within 10 µm when the target is between 90 µm and 120 µm to 48% when the target falls below 91 µm.41 Nevertheless, with the proper microkeratome head selection based on central corneal thickness, the majority of lenticules fall below 131 µm.41 The double-pass microkeratome technique consists of an initial “debulking” pass of the microkeratome, usually with a head between 300 and 350 µm (► Fig. 6.1a), and a second microkeratome pass that is adjusted in order to obtain a lenticule thickness between 100 µm to 130 µm (► Fig. 6.1b). The second pass is started 180 degrees away from the first pass and is performed in the opposite direction to obtain a more planar surface. The head used for this step can be selected according to a nomogram developed by Busin (► Table 6.1), targeting a residual bed with a central thickness < 100 µm (► Fig. 6.1c); other algorithms have also been suggested to adjust both the first pass in relationship to the initial corneal thickness and the second pass related to stromal bed thickness prior to the cut.39 Lenticules created with the double-pass technique tend to be more planar, with a more even distribution of thickness from the center to the periphery after the second pass of the microkeratome.39 Pressure in the system must be standardized by raising the infusion bottle to a height of 120 cm above the level of the artificial anterior chamber and then clamping the tubing 50 cm from the entrance. Attention must be given to maintain a uniform, slow movement of the hand-driven microkeratome, requiring 4 to 6 seconds for each dissection, which will produce a planar surface that minimizes interface irregularities and hyperopic shift. ► Fig. 6.2 shows the optical coherence tomography images of donor tissue before microkeratome double-pass tissue preparation for ultrathin DSAEK (► Fig. 6.2a,b), and before regular single-pass microkeratome for DSAEK preparation (► Fig. 6.2c). With the double-pass microkeratome technique, Busin et al reported 100% of the tissues with a postcut thickness < 151 µm, 95.6% < 131 µm, and 78.3% < 101 µm; and only 2.1% of the tissues were lost due to perforation.19 Woodward et al45 obtained 65% of cuts < 100 µm and 92% < 131 µm with the double-pass technique and found no significant difference in rate of perforation when the second pass was done 180 degrees from the first cut (23%) or at the thickest peripheral measurement in the residual bed (29%). Their higher perforation rate compared to the 7% reported by Busin et al19 can be related to the larger chosen head size for the second cut. The role of the femtosecond laser in donor preparation for DSAEK and ultrathin DSAEK lenticules is still being determined. So far, despite the advantage of more accurate determination of depth of cut, and a faster recovery of visual acuity during the immediate postoperative period, femtosecond laser preparation of DSAEK donors has yielded worse long-term postoperative visual acuities and a higher rate of regrafting when compared to microkeratome grafts.46 In vitro studies of DSAEK lenticule surfaces after microkeratome and femtosecond cuts have also shown that the surface is much smoother when the microkeratome is used for the cut47 and that femtosecond cuts have a wavelike or concentric ring configuration of the stromal interface, caused by the applanation of the cornea during the preparation leading to an irregular cut shape due to incomplete cut of the stroma by the femtosecond.48,49 The surgical procedure for ultrathin DSAEK is essentially the same as that for standard DSAEK. Five-millimeter incisions have been shown to have less endothelial cell loss compared to the 3 mm incisions50,51; however, this difference does not appear to be clinically significant postoperatively.50 The Busin glide and injection devices appear to preserve the endothelium better, compared to manual insertion with forceps,52,53 yet this difference does not appear to be clinically relevant postoperatively either.50 The ultrathin DSAEK lenticule slides without difficulties through the 3 mm incision when inserted through a Busin glide; however, this does not necessarily translate into postoperative advantages, and the tissue insertion technique should be selected to maximize the surgeon’s performance and minimize iatrogenic tissue trauma through excessive manipulation. Placing the sutures prior to viscoelastic removal will provide a pressurized eye, which facilitates suture placement to reduce postoperative astigmatism. It is of equal or greater importance to insert the tissue and then tie the sutures prior to unfolding since not doing so increases the chances of tissue expulsion through the open main incision and through which ultrathin lenticules can be expelled more easily. Table 6.1 Proposed nomogram by Busin et al19 for microkeratome head selection after the first microkeratome pass in the double-pass technique
6.1 The Role of Lenticule Thickness in Postoperative Outcomes
6.2 Interface and Lenticule-Related Higher-Order Aberrations
6.3 The Search for the Thinnest Transplantable Lenticule
6.4 Ultrathin DSAEK: Indications
6.5 Donor Tissue Preparation
6.6 Single Microkeratome Pass
6.7 Double Microkeratome Pass
6.8 Femtosecond Laser Ultrathin Lenticule Preparation
6.9 Surgical Procedure
Residual stromal bed (µm) | Head selection for second microkeratome pass (µm) |
< 151 | No second cut |
151–190 | 50 |
191–210 | 90 |
211–230 | 110 |
> 230 | 130 |
Transient interface fluid has been associated with transient interface opacities54; therefore, similar precautions need to be taken when performing ultrathin DSAEK in removing all fluid and ocular viscoelastic devices from the donor–host interface. The use of a venting incision has been controversial due to the increased risk of epithelial downgrowth into the interface. The author uses a LASIK Flap Roller (Lindstrom, Visitec®), but any other blunt instrument can be used to “sweep” the corneal surface prior to lenticule adhesion with full air fill in the anterior chamber in order to remove any fluid present in the interface.
6.10 Postoperative Outcomes
After EK there is a degree of corneal deturgescence, mostly during the first postoperative month.13,25,31,55,56 This causes a decrease in corneal light scatter, especially in subepithelial surface and donor–host interface during the postoperative follow-up, which may be responsible for most of the improvement in visual acuity. Graft thickness has not been directly correlated to the degree of deturgescence; however, the stabilization of postoperative graft thickness is reached earlier with thinner lenticules.38,56
The main difference between outcomes in DSAEK and ultrathin DSAEK seems to lie in the proportion of patients achieving distance corrected visual acuity (DCVA) of 20/20 by 1 year.19,57 By year 2, this gap between both procedures narrows, with visual acuities of 20/20 or better in 34% of DSAEK patients57 compared to 48.8% after ultrathin DSAEK,19 with a trend toward better visual acuity after ultrathin DSAEK.
DMEK shows both superior results and faster recovery than DSAEK: by postoperative month 6, 42% of patients have DCVA of 20/20 or better, 72% of patients reach 20/25 or better, and 91% 20/30 or better. By the first year, the proportion of patients that reach DCVA of 20/20 remains stable, and 80% and 98% reach DCVA of 20/25 and 20/30, respectively.58 To date there have been no direct comparisons between DMEK and ultrathin DSAEK.
The known postoperative hyperopic shift occurs after all cases of EK. In DSAEK cases, this shift averages + 1.1 D (range + 0.7 to + 1.5D).33 When ultrathin lenticules are used, this shift decreases to + 0.85 D (range – 4.50 D to + 3.25 D) when a single-microkeratome pass is used for preparation56 and + 0.78 D (range + 0 D to + 3.25 D) when the double-pass microkeratome technique is used.19 After DMEK there is still a hyperopic shift, but it is considerably lower at + 0.24 D (- 1.50 D to + 2.25 D).58
6.11 Postoperative Complications
Some known complications from DSAEK, such as intraocular pressure elevation (from pupillary block or from steroid response) and interface blood or haze can also occur in ultrathin DSAEK in a comparable rate.33,56 Endothelial cell loss also seems to be comparable between these two procedures (42% for DSAEK and 35% for ultrathin DSAEK by 12 months),19,33,59 in contrast with endothelial cell loss after DMEK in the same period of time, which has been reported as 19%.59 Graft dislocation, detachment, and rejection rates are somewhat different between DSAEK, ultrathin DSAEK, and DMEK.
6.12 Detachment and Dislocation
There seems to be a trend toward higher rates of detachment or dislocation in thicker grafts and older recipients.28,60 The rates of postoperative graft detachment and dislocation after ultrathin DSAEK19,56 are lower than those reported in large DSAEK case series33,60,61 and considerably lower than those reported for DMEK, which is probably related to the surgeon’s learning curve18,59,62,63 and the large proportion of partial detachments in DMEK cases.59,63,64,65 The vast majority of these detachments resolved after an air reinjection (or “rebubbling”) procedure, which is less commonly required after ultrathin DSAEK19 than after DSAEK33,66 and DMEK.58,65,66
6.13 Primary Graft Failure
There are concerns of increased difficulty in manipulation of ultrathin compared to standard lenticules; however, this has not translated to an increased rate or primary graft failure from excessive tissue manipulation in reported series of ultrathin DSAEK cases19,56 compared to standard DSAEK.33,67 The increased rate of 4% after DMEK64 is likely related to iatrogenic failure due to complexity in tissue manipulation since the main reported cause of primary graft failure in DMEK patients is intraoperative level of difficulty.68 The histopathological findings in removed DSAEK lenticules show in most primary graft failure cases a significant degree of endothelial cell attenuation, evidence of retained material either on the interface stromal side or in the endothelial side, and retained host Descemet membrane or presence of full-thickness cornea from eccentric trephination,69,70 and electronic microscopy of DMEK failed grafts most commonly shows a decreased density of endothelial cells and thickened Descemet membrane with diffuse abnormal collagen inclusions. This demonstrates that the main causes of primary graft failure in all EK cases are related to the procedure and tissue manipulation and a possible preexisting endothelial cell dysfunction prior to transplantation.68
