Surgical management of Fuchs endothelial corneal dystrophy (FECD) continues to advance at a rapid pace. Transition to lamellar surgery is now evolving into a search for more minimally invasive techniques. Several centers, including our own, have reported success with graftless procedures as a primary intervention.1–8 This calls into question the need for endothelial transplantation in every case of symptomatic Fuchs dystrophy and invites the possibility of more inventive use of available tissue for transplantation. In addition, medical therapies for the condition are the subject of intense research. Notable among these is the use of topical Rho-associated kinase (ROCK) inhibitors to aid in corneal clearance.9–16
ROCK is a serine/threonine kinase and an important downstream effector of Rho guanosine triphosphatase, playing a critical role in regulation of contractile tone of smooth muscle tissues.15 Its inhibition as a potential effector of corneal endothelial healing has been pioneered in recent years by Kinoshita et al. Ripasudil hydrochloride hydrate (Glanatec ophthalmic solution 0.4%, Kowa Co Ltd., Nagoya, Japan) is a topical, selective ROCK 1 and 2 inhibitor, which has recently come to the market in Japan for treatment of glaucoma and ocular hypertension. Its safety in phase III clinical trials for ocular hypertension has been demonstrated, as has its ability to accelerate corneal endothelial healing in rabbit models. No adverse alteration of endothelial cell morphology was reported in phase I human trials.9,15,16 Its use for acceleration of corneal endothelial healing is currently off label.
We recently reported our experience of spontaneous corneal clearance after primary descemetorhexis in Fuchs dystrophy.1 Clearance was observed by 1 month with improvement in best-corrected vision from 20/40 to 20/20. Subjectively, the patient reported satisfaction with the procedure and resolution of blur and debilitating glare. There have also been recent reports of mixed or poor results illustrating the potential for failure of this technique.17–19 Caution has been suggested until we are better able to identify successful candidates, or medically rescue any failure.
To this end, we report the outcomes of our surgical trial of descemetorhexis for Fuchs dystrophy. Clinical results were correlated with patient characteristics to attempt to define the target population for the procedure. The study protocol was designed for close observation and salvage with endothelial keratoplasty.
During the course of the study, phase I corneal endothelial safety data on ripasudil in human subjects became available.9 The potential for this as salvage therapy with prevention of a corneal graft was reviewed.
The study was a prospective, interventional case series of primary descemetorhexis for FECD. The study received ethical approval from The University of New South Wales Human Research Ethics Committee. Candidates were enrolled from a tertiary referral corneal clinic at Sydney Eye Hospital, Sydney, Australia. Candidates were chosen based on a clinical profile resembling our successful pilot case. The key criteria were visual symptoms believed to be derived from central guttae rather than corneal edema, the presence of a healthy peripheral endothelial cell population, and symptoms troubling enough to otherwise prompt surgical intervention in the form of endothelial keratoplasty. Full inclusion and exclusion criteria are listed below (Table 1).
At baseline, the complete ocular history was analyzed and dilated fundus examination was performed, with all causes of reduced vision in each eye recorded. FECD was assigned a Krachmer grade by a study investigator, and the presence or absence of diurnal visual fluctuation was noted. Slit-lamp photography and Pentacam (Oculus GmbH, Wetzlar, Germany) corneal tomography were performed. In vivo white-light confocal microscopy (Confoscan 4; Nidek Technologies, Padova, Italy) was performed to acquire en face images of the corneal endothelium with 500× magnification and a lateral resolution of 0.6 mm/pixel. Several images were taken centrally and in the superior periphery. Manifest and autorefraction were recorded, as was contrast sensitivity measured using the CSV 1000 (Vector Vision, Dayton, OH). Anterior segment OCT images were obtained with the Casia SS-1000 (Tomey Corp, Nagoya, Japan).
Candidates underwent central descemetorhexis under viscoelastic, with the intention being to clear the visual axis, but not exceed 4 mm in rhexis diameter. After surgery, topical therapy was started in the form of ketorolac four times daily (qid), chloramphenicol qid, dexamethasone qid, and hypertonic saline (sodium chloride 5%) 6 times a day. Dexamethasone and chloramphenicol were ceased after 1 week, with the other 2 agents continued until corneal clearance.
Baseline measurements were repeated monthly until corneal clearance, then at intervals of 6 months. Nonclearance by 2 months activated more frequent observation and consideration of endothelial keratoplasty.
With mounting laboratory and clinical evidence supporting the use of topical ROCK inhibitors as a method of facilitating human corneal endothelial healing, the potential for the use of these agents to avoid corneal transplantation was considered. After review by the ethics and local drug committees, and discussion with independent corneal specialist members of the study's data safety monitoring board, a decision to use the agents was made, and 3 subjects were withdrawn from the study to receive salvage therapy with the topical agents.
Salvage treatment consisted initially of a 2-week course of ROCK inhibitor administered as topical therapy as per the protocol described by Kinoshita for Y-27632.12 Graft tissue and theater time were allocated if no response was observed by day 14. Two preparations of ROCK inhibitor were obtained for trial: Y-27632 at 10 mM and ripasudil hydrochloride hydrate (Glanatec ophthalmic solution 0.4%).
A multiple linear regression model was used with time to clearance as the dependent variable and age, central cell count, and superior cell count as the independent variables. Statistical analysis of patient factors (age and central and superior endothelial cell counts) was attempted to determine their relevance in predicting time to clearance. The change in preoperative and postoperative best-corrected vision was assessed for significance through a paired t test.
Twelve eyes of 11 patients were enrolled in the study. Follow up ranges from 4 to 30 months. Results are summarized in Table 2; all data listed are as obtained at last follow-up.
Spontaneous corneal clearance was achieved in 9 of the 12 eyes (Fig. 1). Two eyes cleared by 2 months, 5 eyes by 3 months, and 2 eyes by 6 months. In the slowly healing eyes, continuous slow progress was observed. In 3 eyes (cases 7, 10, and 12), clearance was observed to halt, and there was no progress for several weeks, activating salvage treatment.
Preoperative best-corrected visual acuity (BCVA) ranged from 0.0 to 0.6 (logMAR) with a mean of 0.26. Postoperatively, the BCVA in cases achieving corneal clearance ranged from 0.0 to 0.5 (logMAR) with a mean acuity of 0.125. The average gain in best spectacle corrected visual acuity (BSCVA) among corneas achieving clearance was 0.16 (logMAR). This improvement was statistically significant (P = 0.0388). The functional visual acuity, as measured by the CSV 1000 contrast sensitivity assessment instrument, ranged from 0.4 to 0.7 (logMAR) with a mean of 0.52 preoperatively. Postoperatively, this improved to 0.2 to 0.7 (logMAR) with a mean of 0.38.
The mean spherical equivalent remained constant at +0.6 D preoperatively and +0.65 D postoperatively. Central corneal thickness improved from a mean of 615 μm preoperatively to 577 μm postoperatively. In phakic patients, no change to lens density was noted [Lens Opacities Classification System (LOCS)] during the follow-up period. No patient progressed to cataract surgery within the follow-up period. No eye within the study lost lines of BCVA. Within this small sample size, neither age nor endothelial cell counts in central or superior location were found to be significantly associated with time to clearance.
In case 7, corneal clearance was not observed at month 4. Partial detachment of the descemetorhexis edge was observed (Fig. 2). Reattachment was attempted with return to theater and removal of the fibrotic rim, with placement of an air bubble. One month after this second surgery, no improvement was observed. A trial of ROCK inhibitor was then attempted. Topical therapy was started with 10 mM Y-27632. Drops were administered 6 times daily for 2 weeks. After 2 weeks, no measurable improvement was observed clinically or on confocal microscopy, necessitating progression to endothelial keratoplasty. At last follow-up, BSCVA was 20/25 (logMAR 0.1).
In case 10, minimal clearance was observed at 2 months. A small area of detachment of Descemet membrane was noted; this was reattached with air in return to theater. This maneuver failed to influence corneal clearance. At 3 months, no further clearance was observed, and topical therapy with ripasudil was started 6 times per day. Ten days after initiation of ripasudil, complete corneal clearance was achieved with improvement of BSCVA above baseline (Fig. 3).
In case 12, corneal clearance was observed to stall at week 6, with no improvement at week 9. Topical therapy with ripasudil 6 times per day was started with complete clearance by day 14 of treatment (Fig. 3).
In case 9, central corneal clearance was observed by month 2, but a small patch of microcystic edema persisted outside the visual axis, within the stripped area at month 5. Based on the success in other cases, the patient was offered ripasudil. This was started 6 times a day. At 2 weeks, the small patch of edema had fully cleared.
Cell Counts and Morphology
In cases with over 12 months of follow-up, superior counts decreased by 33% (case 1), 48% (case 2), 3% (case 3), 41% (case 4), and 45% (case 5). Multiple superior zones were counted in case 3 to confirm this outlying value. Confocal scanning repeatedly demonstrated the presence of bilobed, hyperreflective nuclei within a single cell in the superior corneal zone (Fig. 4).
Three patients had partial Descemet detachment observed during follow-up. Two of these cases were nonresponders. Rebubbling was attempted in each case, with no successful reattachment achieved. In one case with detachment, complete clearance was still observed, in the presence of a small segment of Descemet detachment. Three patients had formation of endothelial nodules, which were observed to fade but not disappear with time (Fig. 5).
In select patients with visual degradation from guttae in FECD, we believe that descemetorhexis without transplantation may be a viable intervention. Our prospective case series details success (defined as corneal clearance and visual improvement) in 11 of 12 enrolled patients, 9 without any use of ROCK inhibitor. In one case, a subsequent endothelial keratoplasty was needed, with a good visual outcome. Our series also provided promise that ripasudil may represent an accelerator of endothelial healing, which opens a door to possible wider application of this technique. Caution is needed in the face of these early results.
The ability of corneal guttae to degrade vision has been well studied and recently linked to increased forward light scatter.20 Guttae are excrescences of Descemet membrane, with their simplest method of removal being descemetorhexis, a surgical step with which most corneal surgeons are familiar. The necessity for grafting in every case after this step has been performed is now under debate.
Our inclusion criteria outline our selection method, but it is important to highlight that all patients enrolled were requesting intervention for symptoms related to the presence of corneal guttae. Chief among these were glare and blur. It is equally important to note that the presence of low endothelial cell counts or established edema were exclusion criteria in our study. We do not regard descemetorhexis in isolation as a restorative procedure and it should not be considered as therapy for cases of advanced FECD or bullous keratopathy. Indeed, both Descemet membrane endothelial transfer (DMET) series from Melles and ROCK inhibitor work from Kinoshita demonstrated failure in bullous keratopathy cases.4,12 It is likely that some physical reserve is exhausted in these situations beyond all nongraft measures. The observation of possible mitotic bodies on confocal scanning may indicate that some cell division is stimulated by the removal of cell–cell contact inhibition. The downward trend in cell counts, however, would suggest that it is still cell migration that is chiefly responsible for healing in these cases.
The pattern of healing after descemetorhexis was variable. We would agree with the terminology introduced by Colby that there are fast responders, slow responders, and nonresponders to the technique.2 To our surprise, in our small cohort, we could not find a correlation of healing time with age or peripheral endothelial cell counts. From previous reports, it would seem that without the aid of ripasudil, the most significant determinant of corneal clearance is likely to be the rhexis size. One group with a size of 8 mm reported low clearance rates, and 2 isolated reports using 5- and 6-mm sizes also reported failure.17–19 In contrast to using a rhexis size of 4 mm or less, Colby's and our group reported similar spontaneous clearance rates of 75%.1,2 The slowest responders in the series cleared by 6 months. In these cases, lines of BCVA, although not lost, were not gained, indicating a possible need to accelerate corneal clearance if possible for maximal visual improvement. The etiology of the nodules observed in some cases is uncertain but could represent keratocyte-induced collagen formation as a scarring response to Descemet trauma. A mild hyperopic shift was observed in most cases, most likely because of central “thinning” induced by an improved endothelial pump function after guttae removal.
The ability of ripasudil to salvage failing cases was startling, and suggests that cases previously considered borderline may prove to be candidates for descemetorhexis. Before starting this therapy, we firmly believed that a graft would be required in the failing cases, and indeed, theater time and tissue were allocated. It is not clear why ripasudil was effective in 2 cases and compounded Y-27632 was not in the one case in which it was trialed. This may have been due to the differing inhibitions of ROCK 1 and ROCK 2 by the 2 products. We now believe that in the case proceeding to endothelial keratoplasty, the use of ripasudil would possibly have avoided this step.
Ripasudil hydrochloride hydrate 0.4% eye drops were approved for the treatment of glaucoma and ocular hypertension in Japan in 2014. Ripasudil is highly selective for both ROCK 1 and ROCK 2, with minimal off-target effects anticipated as a result. In phase III clinical trials for ocular hypertension, the main adverse event recorded was conjunctival hyperemia in 73.6%, which was transient and did not lead to cessation of treatment in any case.15 Transient formation of guttae-like bodies with morphologic changes of the endothelium has also been reported 1.5 hours after administration in 6 human subjects by Nakagawa et al.9 These changes were found to be reversible with no long-term consequence to the endothelium observed. Although our use of the agent is for a separate indication, the demonstration of safety over long-term dosing periods in glaucoma trials is encouraging. By comparison, there is no further need for use of the agent once corneal clearance has been achieved, which occurred in our subjects within 2 weeks. Subsequent confocal scanning demonstrated a stable hexagonal cell profile, with no fibroblastic transformation.
Descemetorhexis in isolation can rehabilitate vision in Fuchs dystrophy, and it is our impression that it does have a valid role. This may prove to be only a temporizing measure with grafting eventually required. However, with the use of ripasudil, it seems that this role may be expanded, possibly allowing for larger rhexis sizes or application to cases previously felt borderline. Caution must, of course, always be exercised in the face of new drug therapies. Although phase III clinical data are available for ripasudil in the setting of glaucoma, this is still to date only in 1 country. It is our understanding that Kinoshita's group is currently working to reposition this drug for the treatment of endothelial injury.9 Our independent experience would strongly support this strategy. More long-term safety data in this setting are now required, but based on our impressions from this trial, there is tremendous promise that we are moving as a community toward more “non-graft” options for FECD.
1. Moloney G, Chan UT, Hamilton A, et al. Descemetorhexis
for Fuchs' dystrophy. Can J Ophthalmol. 2015;50:68–72.
2. Borkar D, Veldman P, Colby KA. Treatment of Fuchs endothelial dystrophy by Descemet stripping without endothelial keratoplasty. Cornea. 2016;35:1267–1273.
3. Ziaei M, Barsam A, Mearza AA. Spontaneous corneal clearance despite graft removal in Descemet stripping endothelial keratoplasty in Fuchs endothelial dystrophy. Cornea. 2013;32:e164–e166.
4. Dirisamer M, Ham L, Dapena I, et al. Descemet membrane endothelial transfer: “free-floating” donor Descemet implantation as a potential alternative to “keratoplasty”. Cornea. 2012;31:194–197.
5. Shah RD, Randleman JB, Grossniklaus HE. Spontaneous corneal clearing after Descemet's stripping without endothelial replacement. Ophthalmology. 2012;119:256–260.
6. Lam FC, Bruinsma M, Melles G. Descemet membrane endothelial transfer. Curr Opin Ophthalmol. 2014;25:353–357.
7. Bleyen I, Saelens I, van Dooren B, et al. Spontaneous corneal clearing after Descemet's stripping. Ophthalmology. 2013;120:215.
8. Satué Palacián M, Sánchez Pérez A, Idoipe Corta M, et al. Descemetorrexis y aclaramiento corneal: una nueva perspectiva en el tratamiento de las enfermedades del endotelio [in Spanish]. Arch Soc Esp Oftalmol. 2014;89:1–3.
9. Nakagawa H, Koizumi N, Okumura N, et al. Morphological changes of human corneal endothelial cells after Rho-associated kinase inhibitor eye drop (ripasudil
) administration: a prospective open-label clinical study. PLoS One 2015;10:e0136802.
10. Okumura N, Koizumi N, Ueno M, et al. Enhancement of corneal endothelium wound healing by Rho-associated kinase (ROCK) inhibitor eye drops. Br J Ophthalmol. 2011;95:1006–1009.
11. Koizumi N, Okumura N, Ueno M, et al. Rho-associated kinase inhibitor eye drop treatment as a possible medical treatment for Fuchs corneal dystrophy. Cornea. 2013;32:1167–1170.
12. Okumura N, Koizumi N, Kay EP, et al. The ROCK inhibitor eye drop accelerates corneal endothelium wound healing. Invest Ophthalmol Vis Sci. 2013;54:2493–2502.
13. Koizumi N, Okumura N, Ueno M, et al. New therapeutic modality for corneal endothelial disease using Rho-associated kinase inhibitor eye drops. Cornea. 2014;33(suppl 11):S25–S31.
14. Okumura N, Inoue R, Okazaki Y, et al. Effect of the Rho kinase inhibitor Y-27632 on corneal endothelial wound healing. Invest Ophthalmol Vis Sci. 2015;56:6067–6074.
15. Garnock-Jones KP. Ripasudil
: first global approval. Drugs. 2014;74:2211–2215.
16. Okumura N, Okazaki Y, Inoue R, et al. Effect of the Rho-associated kinase inhibitor eye drop (ripasudil
) on corneal endothelial wound healing. Invest Ophthalmol Vis Sci. 2016;57:1284–1292.
17. Galvis V, Tello A, Berrospi RD, et al. Descemetorhexis
without endothelial graft in Fuchs dystrophy
. Cornea. 2016;35:e26–e28.
18. Koenig SB. Planned Descemetorhexis
without endothelial keratoplasty in eyes with Fuchs corneal endothelial dystrophy. Cornea. 2015;34:1149–1151.
19. Arbelaez JG, Price MO, Price FW Jr. Long-term follow-up and complications of stripping Descemet membrane without placement of graft in eyes with Fuchs endothelial dystrophy. Cornea. 2014;33:1295–1299.
20. Oie Y, Watanabe S, Nishida K. Evaluation of visual quality in patients with Fuchs endothelial corneal dystrophy. Cornea. 2016;35:S55–S58.