To describe various endothelial migration healing patterns after Descemet membrane endothelial keratoplasty (DMEK), and to determine the contribution of the donor and host endothelium in the clearance of a transplanted cornea.
Nonrandomized, prospective clinical study.
In a total of 150 consecutive eyes that underwent DMEK (ie, transplantation of an isolated Descemet graft, for Fuchs endothelial dystrophy), re-endothelialization patterns were studied. Of these eyes, 36 showed a “stromal gap” between the “descemetorhexis edge” and the graft, or (partial) graft detachment. Endothelialization patterns of the host posterior stroma were documented at 1, 3, 6, 9, 12, and 24 months after surgery with Pentacam imaging, specular microscopy, optical coherence tomography, confocal microscopy, and slit-lamp biomicroscopy.
Complete corneal clearance was seen in 28 of 36 eyes (78%) with a stromal gap, or (partial) detachment, progressing from the periphery toward the center; and 27 of 34 eyes (79%) with normal visual potential reached a visual acuity of ≥20/40 (≥0.5) or better. In 3 eyes that had the Descemet graft implanted upside-down, a “reversed corneal clearance pattern” was observed (ie, persistent edema where the graft was attached), while the area overlying the detachment cleared. One case that had a “descemetorhexis” performed without endothelial graft implantation showed persistent stromal edema.
The presence of donor endothelium in the recipient anterior chamber may be required for endothelial migration and/or recovery of corneal clarity. Re-endothelialization may be associated with massive endothelial migration and some form of cell signaling to draw donor endothelial cells toward the recipient posterior stroma (“homing”).
The endothelial cell layer of the human cornea may have limited regenerative capacity. Corneal endothelial disorders requiring treatment, most commonly Fuchs endothelial dystrophy and bullous keratopathy, are therefore routinely managed with a corneal transplant, to replace the recipient endothelium by donor tissue. Penetrating keratoplasty has been the gold standard until various new concepts for endothelial keratoplasty were introduced in the last decade, popularized as “deep lamellar endothelial keratoplasty” (DLEK) and (femtosecond) Descemet stripping (automated) endothelial keratoplasty (DSEK/DSAEK/FS-DSEK). Recently, we described a technique for selective transplantation of an isolated Descemet membrane carrying its endothelium, referred to as Descemet membrane endothelial keratoplasty (DMEK). All of these techniques share the basic concept that the donor endothelium restores a normal, balanced stromal hydration, associated with normal corneal transparancy, and thus a recovery of the visual acuity of the recipient eye.
Recently, we reported unexpected corneal clearance with visual recovery up to 20/28 (0.7) Snellen and 20/20 (1.0) Snellen in 2 DMEK eyes that showed (near) complete graft detachment in the early postoperative phase. Slit-lamp observation suggested that cellular repopulation of the host posterior stroma originated from the recipient peripheral endothelium: in the presence of a clearly detached Descemet graft, both corneas cleared from the periphery toward the center, and re-endothelialization was seen to begin in the corneal quadrant opposite to the position of the graft. This clinical finding may not agree with an incapacity of the host endothelium to regenerate, and would potentially challenge the entire concept of a “keratoplasty” as the preferred treatment method in the management of corneal endothelial disorders.
Hence, in the current study we categorized various patterns of “endothelialization” as observed in a series of 36 eyes from a total of 150 DMEK cases to determine how the donor or host endothelium may be involved in the clearance of a transplanted cornea, and to define prerequisites for potential surgical approaches as alternatives to “keratoplasty” in the management of corneal endothelial disorders.
Materials and Methods
A total of 150 eyes of 130 patients enrolled in our prospective study. All eyes underwent DMEK for Fuchs endothelial dystrophy, as previously described. In short, from corneoscleral buttons stored by organ culture at 31 C for 1 week, Descemet membrane (DM) was stripped off, so that a 9.5-mm-diameter flap of posterior DM with its endothelial monolayer was obtained. Due to the elastic properties of the membrane, a “Descemet roll” formed spontaneously, with the endothelium at the outer side. Each Descemet roll was then stored “free-floating” in organ culture medium until the time of transplantation.
In recipient eyes, a 9.0-mm-diameter “descemetorhexis” was created, and the central portion of DM was removed from the eye. The donor Descemet roll was stained with a 0.06% trypan blue solution (VisionBlue, D.O.R.C. International, Zuidland, The Netherlands), and sucked into a custom-made injector (D.O.R.C. International), to transfer the tissue from the culture medium vial to the anterior chamber. Using the injector, the donor Descemet roll was inserted into the anterior chamber and the graft was oriented endothelial side down (donor DM facing recipient posterior stroma) onto the recipient posterior stroma by careful, indirect manipulation of the tissue with air and fluid. The anterior chamber was completely filled with air for 45 to 60 minutes, followed by an air-liquid exchange to pressurize the eye. All surgeries were recorded on DVD (Pioneer DVR-RT601H-S, Tokyo, Japan).
DMEK eyes were examined before and at 1, 3, 6, 9, 12, and 24 months after surgery with Pentacam imaging (Oculus, Wetzlar, Germany), noncontact specular microscopy (SP3000; Topcon Medical Europe BV, Capelle a/d IJssel, The Netherlands), anterior segment optical coherence tomography (Heidelberg Engineering GmbH, Heidelberg, Germany), confocal microscopy, and slit-lamp biomicroscopy (Topcon Medical Europe BV).
Automated specular microscopy was used to document the presence of endothelial cells and the endothelial cell density (ECD) in the area of a “gap” (the exposed recipient posterior stroma in between the edge of the “descemetorhexis” and the Descemet graft), or underneath a detached Descemet graft (the exposed recipient posterior stroma underneath a free-donor-Descemet flap “hanging” in the recipient anterior chamber). In addition, confocal microscopy was used to visualize the cells. To determine the functionality of the endothelial cells, pachymetry measurements over the gap or the area with a detachment were compared to the opposite or adjacent corneal quadrants.
For statistical analysis, the 36 eyes showing graft decentration or (partial) graft detachment were divided into 4 groups that showed distinctive patterns of corneal clearance (Group I: graft decentration; Group II: partial graft detachment; Group III: graft positioned upside-down; Group IV: complete graft detachment or graft absence). One-way analysis of variance was used to compare the groups (Groups I-IV) to the control group of 106 eyes with a well-centered and completely attached Descemet graft in terms of central corneal ECD and central corneal pachymetry. Contrasts (simple effects) were used to examine whether each group differed significantly from the control group. For comparison of central corneal pachymetry, only Groups I, IIa, and III could be included in the comparisons because of violation of the homogeneity of variance assumption in Groups IIb and IV.
Repeated-measures MANOVA was used to assess the difference between central ECD and peripheral ECD adjacent to the Descemet graft, between Groups I, IIa, and IIb. Contrasts (simple effects) were used to examine in which group(s) central corneal ECD differed significantly from peripheral ECD adjacent to the Descemet graft.
All analyses were performed with SPSS software (version 17.0; SPSS, Inc, Chicago, Illinois, USA), using an alpha level equal to .05; a P value less than .05 was considered statistically significant. Bonferroni correction was used to protect against alpha inflation.
From a total of 150 DMEK eyes, 106 revealed no specific information relevant to the pattern of endothelialization; that is, in the presence of an attached Descemet graft, the transplanted cornea showed a “normal” corneal clearance within 1 to 3 months. Of the remaining 44 eyes, 8 eyes (Cases 37-44; Table ) were excluded from analysis because a secondary DSEK was performed within 3 months after the initial DMEK for graft detachment, so that the postoperative time interval was too short to study endothelial migration. Hence, 36 eyes were available for monitoring the endothelialization of the transplanted cornea and/or recovery of corneal transparency and normal pachymetry. The latter eyes could be categorized in 4 groups (Groups I-IV; Table , and Figure 1 ), as described in detail below.
|Case # (Surgery #)||BCVA 6-12 Months||Status at 6-12 Months||Central Corneal ECD (cells/mm 2 )||Peripheral ECD Adjacent to DM Graft (cells/mm 2 )||Central Corneal Pachymetry (μm)||Pachymetry Over Gap Compared to Adjacent Quadrant(s) (μm)|
|GROUP I: >2 mm Stromal Gap Between Descemet Graft and Recipient Rim Descemetorrhexis|
|1 (18)||20/18 (1.2)||Clear cornea including stromal gap||1700||1180||553||733 − 663 = 70 (+10.6%)|
|2 (24)||20/23 (0.9)||Clear cornea including stromal gap||1140||n.a.||642||918 − 818 = 100 (+12.2%)|
|3 (25)||20/20 (1.0)||Clear cornea including stromal gap||2380||1350||542||706 − 766 = −60 (−7.8%)|
|4 (34)||20/20 (1.0)||Clear cornea including stromal gap||550||n.a.||498||709 − 623 = 86 (+13.8%)|
|5 (37)||20/25 (0.8)||Clear cornea including stromal gap||790||650||600||781 − 675 = 106 (+15.7%)|
|6 (44)||20/20 (1.0)||Clear cornea including stromal gap||2620||490||550||933 − 648 = 285 (+44.0%)|
|7 (130)||20/18 (1.2)||Clear cornea including stromal gap||1800||n.a.||541||707 − 737 = −30 (−4.0%)|
|1568 (± 781) n=7||920 (± 410) n=4||561 (± 47) n=7||+12.1% n=7|
|Compared to control group: P = .3216||Compared to corneal center: P = .0011||Compared to control group: P = .0573||—|
|GROUP IIa: <30% Detachment Descemet Graft|
|8 (15)||20/100 (0.2)||Central wrinkle; clear cornea despite persistent detachment||810||1150||650||1004 − 685 = 319 (+46%)|
|9 (20)||20/25 (0.8)||Clear cornea despite persistent detachment||900||870||492||713 − 652 = 61 (+9.4%)|
|10 (26)||20/30 (0.6)||Clear cornea despite persistent detachment||1820||1560||629||766 − 638 = 128 (+20.1%)|
|11 (47)||20/23 (0.9)||Clear cornea despite persistent detachment||2350||1730||551||722 − 728 = −6 (−0.8%)|
|12 (67)||20/25 (0.8)||Spontaneous reattachment at 6 months; clear cornea||1880||n.a.||466||736 − 748 = −12 (−1.6%)|
|13 (71)||[20/50 (0.4)]||Subnormal BCVA attributed to high myopia; clear cornea despite persistent detachment||2480||1820||489||624 − 697 = −73 (−10.5%)|
|14 (80)||20/18 (1.2)||Clear cornea despite persistent detachment||1090||720||496||779 − 626 = 153 (+24.4%)|
|15 (82)||20/20 (1.0)||Clear cornea despite persistent detachment||1730||600||602||944 − 762 = 182 (+23.9%)|
|16 (93)||20/28 (0.7)||Clear cornea despite persistent detachment||1220||1140||523||814 − 800 = 14 (+1.8%)|
|17 (103)||20/18 (1.2)||Clear cornea with small detachment superior||1600||490||559||851 − 680 = 171 (+25.2%)|
|18 (105)||20/18 (1.2)||Clear cornea despite persistent detachment||1860||n.a.||540||795 − 712 = 83 (+11.7%)|
|19 (106)||20/20 (1.0)||Clear cornea despite persistent detachment||1050||640||451||720 − 802 = −82 (−10.2%)|
|20 (109)||20/30 (0.6)||Clear cornea despite persistent detachment||760||350||522||687 − 712 = −25 (−3.5%)|
|21 (110)||20/20 (1.0)||Clear cornea despite persistent detachment||1570||1380||522||647 − 723 = −76 (−10.5%)|
|22 (122)||20/20 (1.0)||Clear cornea despite persistent detachment||1180||1010||572||821 − 649 = 172 (+26.5%)|
|23 (123)||20/23 (0.9)||Spontaneous reattachment at 6 months; clear cornea||1660||1060||525||706 − 843 = −137 (−16.3%)|
|1498 (± 522) n=16||1037 (± 461) n=14||537 (± 56) n=16||+8.5% n=16|
|Compared to control group: P = .0528||Compared to corneal center: P = .0063||Compared to control group: P = .4648||—|
|GROUP IIb: >30% Detachment Descemet Graft|
|24 (63)||20/40 (0.5)||At 6 months some corneal clearance or re-endothelialization over detachment; secondary DSEK 6 months after DMEK||950||n.a.||949|
|25 (64)||n.a.||Corneal clearance inferior quadrants; secondary DSEK 4 months after DMEK||n.a.||n.a.||833|
|26 (73)||20/150 (0.15)||Clear cornea despite persistent detachment; BCVA not explained by cornea||1950||n.a.||558||845 − 746 = 99 (+13.3%)|
|27 (74)||CF (3/60)||Corneal clearance despite persistent detachment; wrinkled graft in corneal center; secondary DSEK 7 months after DMEK||2270||n.a.||896|
|28 (79)||20/40 (0.5)||Clear cornea despite persistent detachment||2070||n.a.||480||636 − 621 = 15 (+2.4%)|
|29 (97)||20/30 (0.6)||Clear cornea despite persistent detachment||1790||n.a.||780|
|30 (150)||20/28 (0.7)||Clear cornea despite persistent detachment||1800||1200||495||692 − 617 = 75 (+12.2%)|
|1804 (± 456) n=6||—||713 (± 198) n=7||+9.3% n=3|
|Compared to control group: P =.8755||—||Variance too high for statistics||—|
|GROUP III: >30% Detachment Because Descemet Graft Positioned Upside-Down|
|31 (36)||20/28 (0.7)||Clear cornea despite persistent subtotal detachment||340||560||627||725 − 1014 = −289 (−39.9%)|
|32 (38)||20/20 (1.0)||Clear cornea despite persistent subtotal detachment||380||490||569||799 − 972 = −173 (−17.8%)|
|33 (96)||20/100 (0.2)||At 3 months corneal clearance inferior quadrants; re-DMEK 6 months after primary DMEK||530||490||711||663 − 810 = −147 (−18.1%)|
|419 (± 99) n=3||513 (± 40) n=3||636 (± 71) n=3||−25.3% n=3|
|Compared to control group: P = .0000||Compared to corneal center: P = .7483||Compared to control group: P = .0000||—|
|GROUP IV: DMEK Aborted After Descemetorhexis (No Descemet Graft Implanted) or Free-Floating Descemet Graft|
|34 (51)||n.a.||Aborted DMEK after descemetorrhexis. At 4.5 months no detectable corneal clearance or re-endothelialization; secondary DSEK 4.5 months after DMEK||n.a.||906|
|35 (43)||n.a.||Complete detachment. At 4 months no detectable corneal clearance or re-endothelialization; secondary DSEK 4 months after DMEK||n.a.||883|
|36 (72)||n.a.||Complete detachment. At 4.5 months no detectable corneal clearance or re-endothelialization; secondary DSEK 4.5 months after DMEK||n.a.||1175|
|—||988 (162) n=3|
|Variance too high for statistics|
|Cases 37-44: Not Informative Regarding Re-endothelialization: Secondary DSEK <3 Months After Initial DMEK|
Group I: Corneal Clearance in the Presence of a Decentered Descemet Graft
Seven eyes showed a gap of >2 mm over at least 2 clock hours, between the host DM and the Descemet graft (Group I: Cases 1-7; Table ). All of these eyes showed complete clearance of the corneal quadrant(s) not covered by the Descemet graft, within 1 to 3 months ( Figure 1 [Group I] and Figure 2 [Group I]; Table ). At later time intervals, the donor tissue usually integrated so well that only the edges of the graft could be identified. All of these eyes reached a best-corrected visual acuity (BCVA) of 20/25 (0.8) or better.
Group II: Corneal Clearance in the Presence of a Partially Detached Descemet Graft
Sixteen eyes (Group IIa: Cases 8-23; Table ) showed a partially (< 1 / 3 ) detached DMEK graft within the first month after surgery, most commonly with the formation of a (small) peripheral “Descemet roll.” Of these, 2 eyes (Cases 12 and 23) showed a spontaneous adherence of the Descemet graft at the 3- to 6-month follow-up interval. In the remaining 14 eyes, complete corneal clearance of the area peripheral to the Descemet roll (ie, the area not covered by the graft) was observed within the first year, most often in about 3 months ( Figure 1 [Group II] and Figure 2 , Top row [Group II]). Slit-lamp examination showed that the denuded posterior stromal area, devoid of either a recipient or donor DM, had a tendency to clear from the corneal periphery toward the center ( Figure 1 [Group II] and Figure 3 ). All but 2 eyes obtained a BCVA of 20/40 (0.5) or better; 1 eye (Case 8) had a poor visual outcome of 20/100 (0.2) attributed to a central wrinkle in the Descemet graft, and 1 highly myopic eye (Case 13) had a maximal visual potential of about 20/50 (0.4).
An additional 7 eyes (Group IIb: Cases 24-30; Table ) showed a large detachment (> 1 / 3 ) of the Descemet graft. All eyes showed progressive corneal clearance, starting at approximately 3 months after DMEK, which was concurrent with endothelialization of the recipient bare posterior stroma ( Figure 1 [Group II] and Figure 4 ). Despite the extensive detachment (ie, without the graft covering the optical center), 4 eyes still reached a BCVA of 20/40 (0.5) or better at 6 months after surgery. Of the 3 remaining eyes, 1 eye (Case 25) had a secondary DSEK surgery at 4 months (BCVA at 6 months unavailable), 1 eye (Case 27) had a BCVA of counting fingers attributed to extensive wrinkling and/or graft contraction in the area over the pupil, and no explanation could be found for the reduced BCVA in Case 26.