Purpose
To evaluate the scrolling propensity of pre-Descemet endothelial keratoplasty (PDEK) tissue and to compare it with each component of the PDEK tissue, namely the pre-Descemet layer (Dua’s layer [PDL]) and the Descemet membrane (DM).
Design
Experimental laboratory investigation.
Methods
Fourteen human donor sclerocorneal discs in which a type 1 big bubble was obtained by stromal injection of air were studied. The wall of the type 1 big bubble was excised and its grade of scrolling noted. The components of the wall (ie, the DM and PDL) were then separated and the scrolling of each was individually graded. Statistical comparison of the grade of scrolling of each layer and correlation with age was carried out; 25-μm slices of anterior and posterior stroma obtained with the femtosecond laser from 4 control samples were used for comparison. The main outcome measure was the grade of scrolling of PDEK tissue in comparison with PDL and DM.
Results
Mean donor age was 67 years. The mean grade of the scroll formed by PDEK tissue was1.6 compared to 0.64 for PDL alone and 3.6 for DM alone. The difference was statistically significant. No correlation between donor age and degree of scrolling for any of the tissues tested was found.
Conclusion
PDEK tissue scrolls less than DM. PDL scrolls the least. This demonstrates that PDL tissue splints the DM and reduces its scrolling in PDEK. This feature has relevance to tissue preparation, handling, and unscrolling in the eye during endothelial keratoplasty.
Endothelial keratoplasty (EK) is the procedure of choice for indications wherein corneal transplantation is required for endothelial diseases such as Fuchs endothelial dystrophy and pseudophakic bullous keratopathy. Besides providing rapid visual improvement, EK does not compromise the integrity of the eye and does not induce significant astigmatism, thus eliminating almost completely the risks of postoperative graft dehiscence and poor vision secondary to high postoperative astigmatism as seen with penetrating keratoplasty. A relatively reduced incidence of immune-mediated graft rejection is an additional advantage. The common technique employed for EK is Descemet stripping endothelial keratoplasty (DSEK) or an automated version called Descemet stripping automated endothelial keratoplasty (DSAEK), wherein the donor endothelial graft supported by deep stroma of around 100–150 μm is prepared with an automated keratome. Provision of precut donor lenticules by eye banks has made DSAEK a popular choice. Descemet membrane endothelial keratoplasty (DMEK), wherein only the donor Descemet membrane with endothelium (DM) is transplanted, has better visual outcomes. The uptake of DMEK among corneal surgeons has not been as rapid as that of DSEK/DSAEK owing to technical challenges and the rather steep learning curve. DMEK tissue is comparatively difficult to harvest from the donor cornea, is relatively delicate and fragile, scrolls with the endothelium outside, and has to be unscrolled in the eye in order to attach it to the posterior surface of the recipient cornea from which the diseased DM has been stripped off. The challenges encountered in harvesting, handling, and unscrolling of DMEK tissue leads to excessive endothelial cell loss; moreover, the DM is relatively firmly attached to the underlying stroma in younger eyes and the risk of tissue loss during harvesting is reported to be higher in donor eyes below 50 years of age.
Recently, Dua and associates reported the presence of a distinct layer of the deep stroma termed the pre-Descemet layer (Dua’s layer, PDL) which is consistently present in eyes of all ages and separates with the DM as a type 1 big bubble in the vast majority of cases during deep anterior lamellar keratoplasty (DALK) performed by big bubble (BB) technique. Dua and associates first proposed and demonstrated in eye bank eyes that the tissue thus obtained by injecting air in donor eyes, consisting of PDL and DM, could be used for EK, as it offered the advantages of being more robust and scrolled less than DMEK tissue. The term “pre-Descemet endothelial keratoplasty (PDEK)” was introduced to describe the above procedure and the first patients undergoing this procedure were described in 2014.
One of the advantages attributed to PDEK tissue is that it scrolls less, as the PDL affords a splinting effect to the DM, limiting its ability to roll, which in turn makes it easier to unroll in the eye and makes intraoperative handling and centration easier. In this study we analyzed the scrolling propensity of PDL and DM combined and of each layer separately in order to provide evidence in support of the clinical observation.
Methods
This was a prospective, collaborative, ex vivo clinically relevant laboratory investigation on human eye bank donor eyes carried out in 4 centers—namely, the Academic Section of Ophthalmology, Division of Clinical Neuroscience, University of Nottingham, Nottingham, United Kingdom; LaserVision, Nuffield Health Hospital, Guildford, United Kingdom; Department of Ophthalmology and Visual Sciences, University of British Columbia, Vancouver, Canada; and the Department of Ophthalmology, University of Chieti-Pescara, Chieti, Italy. This study adhered to the tenets of the Declaration of Helsinki.
Fourteen human donor eyes were studied. Donor details and causes of death are provided in the Table . Nine sclerocorneal discs were maintained in organ culture in Eagle’s minimum essential medium with 2% fetal bovine serum and 5 in tissue storage medium (Optisol; Chiron Ophthalmics, Inc, Irvine, California, USA) for 2–6 weeks prior to the experiment. All tissue used was from consented donors and released for research by the relevant eye banks.
| Sample | Age | Sex | Cause of Death | Grade of Scroll | ||
|---|---|---|---|---|---|---|
| PDL + DM | PDL | DM | ||||
| E2325 | 57 | M | Cancer | 2 | 0 | 4 |
| E2328 | 71 | M | Stroke | 1 | 0–1 | 3 |
| E2273 | 71 | F | Multiorgan failure | 0–1 | 0–1 | 4 |
| E2277 | 80 | F | Sepsis | 2 | 1 | 4 |
| E2294 | 78 | F | Cancer | 1 | 0–1 | 3 |
| E2271 | 43 | M | ICH | 2 | 1 | 4 |
| E2306 | 70 | M | Liver failure | 1 | 0–1 | 3 |
| E2305 | 75 | M | Renal failure | 2 | 1 | 3 |
| M2247 | 69 | F | COPD | 2 | 0–1 | 4 |
| 0099 | 67 | M | Cancer, lung | 3 | 1 | 4 |
| 0306 OD | 57 | F | Aneurysm | 1 | 0–1 | 4 |
| 0306 OS | 57 | F | Aneurysm | 1 | 0–1 | 3 |
| 0315 | 60 | M | Ischemic heart disease | 3 | 1 | 4 |
| 0334 | 73 | M | Ischemic heart disease | 1 | 0–1 | 0–1 |
The sclerocorneal discs were injected with air as described by Dua and associates. Briefly, the sclerocorneal disc was placed endothelial side up in a Petri dish and a 30 gauge needle attached to a 5 mL syringe was inserted from the scleral rim into the deep corneal stroma up to the center. Counter traction was applied by holding the sclera with a toothed forceps adjacent to the site of needle insertion and an assistant held the scleral rim with a toothed forceps directly opposite the site of needle insertion, to steady the sclerocorneal disc. Air was injected slowly but firmly until a type 1 big bubble was obtained and expanded to its full diameter by continuous injection of air. The walls of the big bubble, consisting of PDL and DM, were excised and the tissue thus obtained was immersed in balanced salt solution (BSS; Alcon Laboratories, Inc, Fort Worth, Texas, USA) in a Petri dish. The diameter of the type 1 big bubble was measured (7.5–8.5mm) and the tissue was excised along the circumference of the big bubble and covered with trypan blue dye (VisionBlue 0.06% Trypan Blue Ophthalmic Solution; DORC Dutch Ophthalmic Research Center International, Zuidland, Netherlands) for 1 minute. The tissue was then slid off the sclerocorneal disc into the Petri dish half-filled with BSS, endothelial side up; completely submerged in BSS; and allowed to rest for 5 minutes. The degree of scrolling of the tissue was observed and graded by the method described by Bennett and associates. They rated (graded) the scroll as 1 when the opposite ends of the scroll did not touch each other, 2 when the opposite ends touched each other, 3 when a complete scroll was formed, and 4 when more than 1 complete scroll formed. In our study, an additional grade, grade 0, was also considered and added to the scale for tissue that did not scroll at all. Grading was performed independently by an observer in Vancouver and another in Nottingham with full concordance.
Thereafter, with 2 pairs of fine nontoothed forceps the edge of the tissue was gently teased to separate DM from PDL. When the DM and PDL edges could be seen separately, they were grasped with the forceps and gently pulled apart. The edge had to be released and regrasped along the circumference of the tissue as the 2 layers were gently teased apart. At final separation, the 2 layers were laid with the apposing surfaces (DM surface of PDL and DM) facing up. The entire maneuver was carried out under BSS. The separated PDL and DM were submerged and allowed to rest for 5 minutes. The degree of scrolling of each layer individually was then regraded as described above.
Samples where a type 2 or mixed bubble formed were discarded.
Statistical Analysis
The nonparametric Friedman test was applied for a statistical comparison of the mean grade of scrolling between the 2 layers individually and the combination of both. Post hoc comparisons were performed using Wilcoxon signed rank test. A generalized linear regression analysis was used to assess the influence of donor age on the grade of scrolling for the 2 layers individually and the combination of both. All statistical tests were evaluated at a 2-sided alpha level of 0.05. Statistical analysis was performed using SPSS 21.0 (IBM, Armonk, New York, USA). A false discovery rate correction for multiplicity was applied to the post hoc and regression analysis independently in order to reduce the risk of a type 2 error (ie, accept a null hypothesis that is actually false or exclude the presence of a statistically significant result when actually present).
For controls, 4 sclerocorneal discs were inverted (epithelial side concave) and mounted on the Ziemer artificial chamber with the endothelial surface up. The DM was peeled off, exposing the PDL. The software on the LDV Z8 femtosecond laser was adjusted to provide horizontal slices of 25 μm of tissue with a 7 mm diameter. The tissue was transferred to the Barron artificial anterior chamber (Katena, Denville, New Jersey, USA) and the slices from the posterior stroma (posterior third) and anterior stroma (anterior third) were dissected and immersed in BSS and then graded for scrolling as above.
Results
The age of donor eyes ranged from 43 to 80 years with a mean of 67 years. The grade of the scroll formed by PDEK tissue (PDL + DM) was between 1 and 2 with a mean of 1.6. In all samples of PDEK tissue and DM the scrolling occurred with the endothelial cells on the outer surface of the scroll. PDL on its own also scrolled such that the edges curled away from the surface that was apposed to the DM; in other words, all 3 types of samples examined, scrolled in the same direction. Trypan blue stain allowed easy visualization of the layers and knowledge of the surfaces that were facing the observer permitted the determination of the direction of scrolling. By agitating the BSS, the scrolls could be made to roll in the solution with the open ends facing up or down, but they always retained their characteristic of the endothelial cells/surface being on the outside. The grade of the scroll of PDL alone was between 0.5 and 1 in all cases, with a mean of 0.64 ( Figure 1 ). The grade of the scroll formed by the DM alone was between 3 and 4 in all cases except 1, where it was 0.5. The mean of DM scrolling was 3.39, but excluding the single outlier it was 3.61. Details of the grade of scroll for the different samples and their respective ages are given in the Table .
