REVIEW

CURRENT
OPINION Descemet’s stripping without endothelial
keratoplasty
Daniel Garcerant a, Nino Hirnschall a,c, Nicholas Toalster a, Meidong Zhu a,d,
Li Wen d, and Gregory Moloney a,b

Purpose of review
To summarize the recent literature regarding descemetorhexis stripping without endothelial keratoplasty
(DWEK), increasingly referred to as Descemet’s stripping only (DSO). To report the characteristic clinical,
confocal and histologic findings associated with this procedure.
Recent findings
Reported clearance rates following DSO range from 63 to 100% in recent series, with variation between
surgical techniques. Topical Rho-kinase inhibitor has been reported as successfully salvaging failing cases.
Its use as an adjuvant to the surgery is gaining widespread adoption with the results of early series now
arriving. Apart from a phenotype of central guttata with clear periphery, patient characteristics which
determine success remain elusive. Surgical factors affecting success are increasingly well understood, with
stromal injury felt to be a retardant to healing. Characteristic clinical signs have been observed and are
described herein. Clinical, confocal and light microscopic images are obtained from patients in clinical
trials of DSO with ripasudil.
Summary
DSO is gaining acceptance as a surgical option for a subset of patients with Fuchs’ Dystrophy. The
addition of Rho-associated kinase inhibitor appears to improve predictability but further results to this effect
must be published and scrutinized.
Keywords
Descemet’s stripping only, descemetorhexis, descemetorhexis stripping without endothelial keratoplasty, Rho-
associated kinase inhibitor, ripasudil

The use of healing by primary intention as a TERMINOLOGY

planned surgical strategy is well accepted in various &
In 2018 Kaufman et al. [20 ] proposed the term
sites in the body. This is primarily in the setting of DWEK (descemetorhexis without endothelial kera-
epithelial closure with reliable mitotic ability. Tra- toplasty) to describe this technique, rightly pointing
ditional understanding of the human corneal endo- out the need for standardized terminology as this
thelium is that this cellular monolayer is locked in operation and its variants are increasingly discussed.
the G1 phase of the cell cycle and unable to replicate Alternative terminology has been proposed by Gor-
in the event of cell loss [1]. This understanding has ovoy, suggesting the term Descemet’s stripping only
been challenged for many years [2,3], with the (DSO) [21]. Both have merits, but also flaws, as the
suggestion being that some mitotic ability is pres-
ent, more so in youth, more so in the corneal
a
periphery [2–5]. Concurrently, the rapid evolution Sydney Eye Hospital, bDiscipline of Ophthalmology, Sydney Medical
in endothelial keratoplasty has led to numerous School, The University of Sydney, Sydney, New South Wales, Australia,
c
VIROS – Vienna Institute for Research in Ocular Surgery, Hanusch
observations of attempted healing in the presence
&
Hospital, Vienna, Austria and dWales Tissue Bank, New South Wales
of an endothelial defect [6–9,10 ,11], with much Organ and Tissue Donation Service, Sydney, New South Wales,
basic science suggesting this process can be acceler- Australia
ated pharmacologically [12–19]. The intersection of Correspondence to Gregory Moloney, Mosman Eye Centre, 1A Effing-
these findings has led to the exploration of removal ham St., Mosman, Sydney 2088, NSW, Australia. Tel: +61 409096605;
of diseased corneal endothelium in Fuchs’ dystro- e-mail: gregorymoloney@yahoo.com.au
phy with planned healing by primary intention as a Curr Opin Ophthalmol 2019, 30:275–285
surgical strategy. DOI:10.1097/ICU.0000000000000579

1040-8738 Copyright ß 2019 Wolters Kluwer Health, Inc. All rights reserved. www.co-ophthalmology.com

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.

Corneal and external disorders

With the evolution of endothelial keratoplasty,

KEY POINTS the deliberate stripping of Descemet’s membrane
 Patient factors affecting clearance rates after DSO/ has become commonplace. The Melles group
DWEK remain difficult to determine. reported clearance of corneas with nonattached
Descemet’s grafts, with this subsequently tried
 Surgical factors affecting clearance rates after DSO/ deliberately in a technique dubbed Descemet’s
DWEK are better defined with a peeling technique &
membrane endothelial transfer [8,9,10 ,11]. Several
preferred over a scoring technique to avoid
stromal trauma. reports emerged at a similar time describing delib-
erate removal of Descemet’s membrane without
 Additional Rho-associated kinase inhibitor has been grafting for Fuchs’ Dystrophy [30,34,35,36 ].
&&

described as salvaging failing cases and in early series Early reports of the success of this strategy were
appears to improve speed of clearance and final
mixed at best (Table 2). Bleyen et al. [34] described
cell counts.
failure in seven of eight cases after stripping Desce-
 Early reports of subsequent graft outcomes if required mets membrane with a diameter of 8 mm. Price [35]
are encouraging. reported mixed outcomes in a series of three cases
with a 6–6.5-mm Descemetorhexis, two clearing with
one case having residual edema. Concerns were held
surgery is still a keratoplasty – a term that is not in all cases regarding postoperative visual quality and
synonymous with grafting. Nor is Descemet’s strip- irregular astigmatism. Koenig [37] also reported non-
ping the only part of this operation, with its success clearance in two cases with a 6 mm descemetorhexis.
increasingly seeming to be enhanced by adjuvant In reports of inadvertent Descemet’s membrane endo-
medical therapies. For the time being the consensus thelial keratoplasty (DMEK) graft detachment and
of opinion may be favoring DSO, which is the term corneal clearance, all published cases eventually pro-
we will use in this review. ceeded to DMEK with re-emergence of corneal edema
after roughly 24 months [38,39]. Of course in these
cases an 8–9-mm descemetorhexis was created.
BACKGROUND In cases with a smaller descemetorhexis better
&&

Although recent publications have given surgeons clearance rates were reported. Iovieno et al. [43 ]
renewed confidence to attempt Descemet’s strip- reported clearance in four of five cases with a 4 mm
ping for Fuchs’ Dystrophy, the idea itself was descemetorhexis but also reported concerns regard-
described in detail decades ago by Paufique [22]. ing irregular astigmatism. In Borkar’s series 10/13
The technique was clearly not widely adopted and cases with a 4 mm descemetorhexis achieved corneal
the reference all but lost, discovered by chance in clearance and in our initial series 9/12 cases with a
&&

the library of the Massachusetts Eye and Ear Infir- 3–4 mm descemetorhexis achieved clearance [36 ].
mary (R Pineda – personal communication). In The issues affecting the surgery at this point in
more recent years the observation of corneal clear- time were the unpredictability of corneal clearance,
ance after iatrogenic trauma was noted by several both who would achieve it, and how long it would
authors (Table 1) [23–33]. take, but also the quality of vision conferred even if

Table 1. Selected case reports of spontaneous corneal clearance following Descemet’s stripping

Case reports Comments

Braunstein (2003) [23] 1/1 Inadvertent removal of DM during cataract surgery

Patel (2004) [24] 1/1
Zvi (2005) [25] 1/1 SF6 (þair)
Pan (2006) [26] 1/1 Air bubbling
Agarwal (2006) [27] 1/1
Srivastava (2010) [28] 1/1 Inadvertant removal of DM, air bubbling
Choo (2010) [29] 1/1
Shah (2012) [30] 1/1 Deliberate DSO as an intervention for FECD
Koenig (2013) [31] 1/1 Maintenance of corneal clarity in FECD patient 16 years post descemetorhexis
Jullienne (2015) [32] 1/1
Sanchez (2017) [33] 1/1

FECD, Fuchs endothelial corneal dystrophy; DM, Descemet’s membrane; DSO, Descemet’s stripping only.

276 www.co-ophthalmology.com Volume 30  Number 4  July 2019

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.

 Descemet’s stri pping without endothelial keratoplasty Garcerant et al.

Table 2. Case series of deliberate Descemet’s stripping in Fuchs’

Case series Outcome (clearance) Descemetorhexis size ROCK inhibitor Y/N

Bleyen (2013) [34] 1/8 8 mm N

Arbelaez (2014) [40] 2/3 6 mm N
Moloney (2015) [41] 2/2 4 mm N
Koenig (2015) [37] 0/2 6 mm N
Borkar (2016) [36 ] 10/13 4 mm, with phaco N
&&

Malyugin (2017) [42 ] 12/19 4 mm, with phaco N

&

Iovieno (2017) [43 ] 4/5 4 mm N

&&

Moloney (2017) [44 ] 9/12 4 mm Y – 2 patients rescued, one patch

&&

of edema resolved with ripasudil

Davies (2018) [45 ] 14/17 4 mm N
&&

Huang (2018) [46 ] 12/12 4 mm, with phaco Y in 3/12 cases

&&

Macsai (2019) [47 ] 9/9 4 mm Y in all cases

&&

ROCK, Rho-associated kinase.

clearance was achieved. Colby coined the useful SURGICAL FACTORS

terminology – ‘fast, slow and nonresponders’ to We have observed that posterior stromal scarring is
&&
the technique [36 ]. Clearly a strategy was required often present in slow or nonresponders at the site of
to preselect the fast responders presurgery or convert recalcitrant edema during the clearing process.
the slow and nonresponders postsurgery. It is rea- Analysis of surgical video demonstrated that this
sonable to assume that a patient’s response to this scarring was associated with stromal scoring in every
surgery is either due to innate, possibly genetic case (Fig. 1). At times this stromal response assumes
characteristics of the anterior segment, or iatrogenic characteristic patterns, with our initial publications
factors at the time of surgery. describing the formation of posterior stromal nod-
ules at the stripped margin. The nodules observed
may represent new collagen laid down in a scarring
PATIENT FACTORS response, or epithelial mesenchymal transforma-
&&
Both our series and that of Davies et al. [45 ] tion of healing endothelial cells, locally acted upon
attempted analysis to determine what predictive by cell signals from injured stroma. Fortunately,
factors might exist for corneal clearance. Both these are seen to fade over time, but are still detect-
groups in small to medium cohorts were unable able at 4 years (Fig. 2).
to identify clinical or demographic factors of signif- The assumption therefore was that surgical
&&
icance. Davies et al. [45 ] noted that in patients trauma to the stroma initiated an unpredictable
undergoing fellow eye treatment, clearance time healing response favoring fibrosis, possibly medi-
was strikingly similar, suggesting a role for patient ated by stromal keratocytes. A surgical technique
factors which thus far remain unclear. A clear link in to avoid stromal contact was felt necessary.
our center between trinucleotide repeat load and Unpublished in-vitro work also demonstrated a
time to clearance has not been established albeit in a more favorable healing response in cases without
small sample size (unpublished data) It is worth surgical trauma to the stroma. Specular microscopic
noting that in an in-vitro analysis of endothelial analysis of the descemetorhexis margin created with
cell migration by Soh et al. [48] identified younger both scoring technique and a peeling technique was
patient age and the presence of an intact Descemet’s analyzed utilizing human donor tissue consented
membrane as factors capable of promoting cell loco- for research and not fit for transplantation. In tech-
&&
motion. Okumura et al. [49 ] also found improved niques in which constant pressure is created during
endothelial cell migration after removal of the cen- Descemet’s scoring (trephination or scoring) we
tral endothelium if Descemet’s membrane was pre- observed the creation of a stromal trench, present-
served rather than stripped. This would suggest that ing a possible challenge to cell locomotion. Loss of
a roughened stromal surface, or bare stromal surface cells in the stripped area was of course anticipated.
is a retardant of endothelial cell migration, and In scoring technique, loss of cells on the host side of
again emphasizes the effect of surgical factors on the wound was also observed, indicating some
the success of this procedure. trauma is sustained on both sides of the stromal

1040-8738 Copyright ß 2019 Wolters Kluwer Health, Inc. All rights reserved. www.co-ophthalmology.com 277

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.

Corneal and external disorders

FIGURE 1. The correlation between surgical trauma and persistent stromal scarring. (a) Surgical still image demonstrating
traction on overlying stroma from grasping forceps, light reflex demonstrates downwards traction. (b) Slit lamp photograph
demonstrating corresponding area of posterior stromal scarring and fibrosis persisting at 12 months postsurgery. (c) Surgical still
image demonstrating scoring of stroma with Sinskey hook. Light reflex demonstrates upwards indentation. (d) Slit lamp
photograph demonstrates curvilinear stromal scar in corresponding area. This was a persistent focus of edema in healing phase.

FIGURE 2. Posterior stromal nodules visible in cases scored with a Sinskey hook at the Descemetorhexis margin. (a) Anterior
segment optical coherence tomography demonstrating hyperreflective nodule at descemetorhexis margin. Nodule is sub
Descemet’s in parts suggesting stromal origin. (b) Corresponding slit lamp image at 3 months post surgery, nodules seen as
whitish opacities at descemetorhexis margin. (c) Same patient, 4 years postsurgery demonstrating flattening of nodule with
reduced reflectivity. (d) Slit lamp image 4 years post surgery demonstrating fading of nodules.

278 www.co-ophthalmology.com Volume 30  Number 4  July 2019

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.

 Descemet’s stri pping without endothelial keratoplasty Garcerant et al.

FIGURE 3. Specular microscopic images obtained with a Nikon CCD camera on light microscope connecting to computer
(Software Tomey EM-1020 version 1.2, Nagoya, Japan). Human donor corneas with Descemetorhexis created via scoring (a
and c) or peeling (b and d). Note creation of a trench with scoring techniques and loss of cells on host side of margin. Note
peeling techniques with preservation of cells to wound edge.

trench (Fig. 3a and c). In descemetorhexis wounds study mean 32 mm) [48]. This is presumably due to
created with a peeling technique, preservation of improved endothelial pump function in the absence
cell morphology was observed to the very margin of of guttata and a thickened Descemet’s membrane.
the wound (Fig. 3b and d). Any corneal procedure producing a central thinned
Based on our experience we strongly recom- zone will act in a similar manner to myopic ablation,
mend using a peeling technique, both to maximize with an optical zone created. We know from decades
cell preservation and migration. This recommenda- of refractive work that a small or decentered optical
&&
tion is reinforced by Davies et al. [45 ], who found zone can be associated with higher order aberra-
that in their cohort all cases who failed to heal after tions. We are currently analysing the effect of cen-
DSO had the common factor of a 360-degree scoring tration on visual outcomes of this procedure and
technique followed by stripping, conversely all cases place emphasis on symmetry and centration as part
stripped without scoring cleared successfully. of surgery. Second if stromal trauma is sustained, the
posterior scarring response described above will cre-
ate an irregular posterior corneal profile, and subse-
VISUAL OUTCOMES AND REFRACTIVE quently an irregular posterior corneal ‘lens’. Third, a
EFFECT more remote possibility – the refractive effect of
Regarding postoperative visual quality, ghosting posterior corneal incisions was noted decades ago
and irregular astigmatism have been observed to by Sato [50]. In his technique, the stroma was delib-
occur in cases despite corneal clearance. We have erately violated to significant depth to create an
three hypotheses to explain this effect. First, it has arcuate keratotomy, it is possible however that
been observed that in cases achieving corneal clear- relaxing incisions in Descemet’s membrane may
ance central corneal thinning may occur (in our last have an astigmatic effect.

1040-8738 Copyright ß 2019 Wolters Kluwer Health, Inc. All rights reserved. www.co-ophthalmology.com 279

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.

Corneal and external disorders

Analysis of visual results in the literature is cur- this will fragment requiring a new tag to be created,
rently complicated by the number of studies combin- this must be done so with minimal stromal trauma.
ing the surgery with cataract extraction, but An effort to maintain Descemetorhexis along the
successful DSO has been shown to be compatible 4 mm mark without leaving any portion inside is
with good visual acuity (all values shown as logMAR recommended. Viscoelastic is then removed with
&&
for better comparison). Borkar et al. [36 ] reported a irrigation/aspiration and main wound is hydrated.
best corrected distance visual acuity (BCVA) between
0.12 and 0.00 logMAR, with two patients excluded
from analysis due to retinopathy. The same group THE ROLE OF ADDITIONAL RHO-
&
[51 ] presented two other cases with 0.00 logMAR ASSOCIATED KINASE INHIBITOR
BCVA after DSO, but another patient with posterior Since their discovery in 1995, the Rho-associated
synechiae and cystoid macular edema with a BCVA of kinase (ROCK) family of enzymes has been the
0.49 logMAR. We presented a BCVA improvement in subject of much research as therapeutic targets.
patients receiving DSO in isolation from 0.26 logMAR The primary effects of Rho proteins are to induce
pre-DSO to 0.13 logMAR post-DSO including those structural change in the internal cell cytoskeleton,
&&
patients with incomplete corneal clearance [44 ]. An inhibiting smooth muscle vasodilation, cellular
interesting recent publication from Huang et al. com- delamination and migration [14]. An in-vitro study
pared visual outcomes of 12 DSO cases with 15 DMEK of Ripasudil’s effect on the trabecular meshwork and
cases. Average time to reach 20/40 vision was longer Schlemm’s canal endothelial cells also showed cyto-
for DSO cases (2.2  2.8 weeks compared with skeletal rearrangement (cell rounding and reduced
7.1  2.7 weeks), but there was no statistical differ- actin bundles). It also decreases ZO-1 cell junctional
ence in final visual acuity (logMAR 0.16  0.09) for complex proteins [52]. Although via kinase path-
DMEK eyes logMAR 0.13  0.10 for DSO eyes ways proliferative effects may be modulated, as cor-
(P ¼ 0.44). Postoperative adverse events were higher neal surgeons our expectation that inhibition of this
in the DMEK cohort, none however resulting in activity will safely induce mitotic activity and
&&
lasting complications [46 ]. This analysis indicates replenish situations of low endothelial cell counts
that visual acuity (only one measure of visual quality) is likely misplaced [12,13,15,16,53–55]. There is
can be equivalent to DMEK after DSO. An important however justified confidence in their ability to pro-
caveat to note is the use of pinhole acuity as a measure mote cell migration in situations where this has
in this series, which may mask some effects of irregu- stalled, with two cases of failing DSO salvaged by
&&
lar astigmatism if present in the DSO cohort. It is ours topical ripasudil use in our center [44 ]. The first
and others impression that patient reported satisfac- double armed prospective study of ROCK inhibitor
tion with the procedure is high, but this is in need of as an adjuvant to DSO has recently been published
proper analysis and publication. by Macsai et al. In a comparison of DSO in patients
with ripasudil postoperatively with DSO without
ripasudil, the ripasudil group had a significantly
CURRENT SURGICAL TECHNIQUE faster recovery and a higher central endothelial cell
The combination of the ‘optical zone’ and ‘stripping count compared with patients without ripasudil
&&
over scoring’ hypotheses has led to the following [47 ]. The trial is small with further analysis now
surgical technique. required but is a positive first step.
Preoperative preparation includes marking the
cornea at the center of the pupil (prior to peribulbar
block) in mesopic conditions to guide centration. NOTEWORTHY CLINICAL SIGNS
The pupil is then dilated to allow for a better red Characteristic clinical signs are associated with this
reflex during descemetorhexis. surgery and the resultant healing response. With the
In the operating theatre a caliper is used to create addition of topical Rho-kinase inhibitor, different
4-mm diameter imprint centered on the previous signs emerge.
mark to delimitate the descemetorhexis. A single
main 2 mm clear corneal incision is created and
the anterior chamber is filled with cohesive viscoelas- Contracting edema/expanding clear zone
tic. A reverse Sinskey hook is positioned in the endo- Initial dense corneal edema contracts as endothelial
thelial edge of the 4-mm circle, then with very gentle migration takes place. The contraction of the area of
pressure over the endothelium, small side to side epithelial edema and formation of a ‘clear zone’
movements create a small Descemet’s membrane between descemetorhexis margin and this edema
tag. The tag is thereupon picked with grasping forceps are measurable signs that allow progress to be
and a Descemet’s tear is propagated in a circle. Often recorded (Fig. 4a).

280 www.co-ophthalmology.com Volume 30  Number 4  July 2019

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.

 Descemet’s stri pping without endothelial keratoplasty Garcerant et al.

FIGURE 4. (a) Slit lamp photograph demonstrating contracting microcystic edema and formation of a clear zone superiorly
and inferiorly (red arrows). (b) Slit lamp photograph demonstrating transformation of microcystic into macrocystic or
‘honeycomb’ edema on treatment with topical Rho-kinase inhibitor.

Failure to form a clear zone or arrest of the epithelium (Fig. 4b). In our recent series, this finding
process of contraction are warning signs for failure. was not recognized at study commencement but is
likely present in all treated cases. The corneal epithe-
lium affected in this way demonstrates a morphologic
Honeycomb edema change in epithelial cells. On confocal microscopy,
Of interest, in cases treated with topical Rho-kinase stratified squamous cells are seen to adopt a more
inhibitor the edema adopts a ‘honeycomb’ appear- spindle shape at the cyst edge, with loss of cell polarity
ance with multiple vacuolated fluid cysts within the and organized stratification (Fig. 5a). The collapse of

FIGURE 5. In-vivo white light confocal microscopy (Confoscan 4; NIDEK Technologies, Padova, Italy). (a) Central corneal
epithelium treated with topical ripasudil. Formation of fluid filled macrocysts with spindle cell morphology at cyst edge. (b)
Peripheral corneal epithelium from same patient demonstrating preservation of normal cell morphology in non edematous area.

1040-8738 Copyright ß 2019 Wolters Kluwer Health, Inc. All rights reserved. www.co-ophthalmology.com 281

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.

Corneal and external disorders

hypothesis is that cell–cell contact inhibition ameli-

orates or negates the effect of topical ROCK inhibition.
Where this contact is broken down at the site of
microcystic edema, ROCK inhibition will produce
further cytoskeletal change.

Pseudoguttata
In early studies of ROCK inhibitor in human patients
‘pseudoguttata’ were noted to arise as a transient dark
spots or ‘cells’ visible on specular microscopy, theo-
rized to be due to disruption of actin microfilament
bundles and impairment of focal adhesion formation
[14]. In vitro analysis of ROCK inhibition in vascular
endothelial cells also demonstrated localized gap for-
mation between cells [56]. It is worth noting that these
have been described as occurring in many scenarios of
FIGURE 6. Light microscopy (H&E stain) of epithelial biopsy ocular inflammation without ROCK inhibition [57].
in the presence of honeycomb edema. Pale cytoplasm in In subjects treated following DSO with supplementary
some cells in presence of edema. Red arrow indicates ROCK inhibitor we have observed similar findings,
spindle cell formation at edge of fluid cyst. with guttata like bodies detectable within the stripped
area of the cornea in the weeks following surgery
the affected cells converts microcystic into macrocys- (Fig. 7c). Confocal microscopy showed these same
tic edema. Histologic analysis failed to demonstrate ‘dark bodies’ seeming to arise at intercellular junctions
any evidence of dysplastic or metaplastic change. rather than within the cytoplasm (Fig. 7a and b). We
(Fig. 6) These findings are reversible within hours of agree that this is likely to represent an interruption of
drop administration and are consistent with the cell adhesion molecules under the influence of ROCK
known primary effect of ROCK inhibition – cytoskel- inhibition and should not be mistaken with Fuchs’
etal reorganization and degradation of intercellular recurrence. It is most visible in the population of cells
junctions. The findings are not found at all in the migrating to cover the descemetorhexis, presumably
non edematous corneal epithelium (Fig. 5b). Our where intercellular junctions are most newly formed.

FIGURE 7. Endoth pseudoguttata. (a and b) In-vivo white light confocal microscopy (Confoscan 4; NIDEK Technologies,
Padova, Italy) images demonstrating round, dark bodies at intercellular junctions (white arrows), correlating to location of
pseudoguttata. (c) Slit lamp photograph of pseudoguttata (orange arrows).

282 www.co-ophthalmology.com Volume 30  Number 4  July 2019

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.

 Descemet’s stri pping without endothelial keratoplasty Garcerant et al.

polymegathism and a larger cell surface area on which

to deposit. Its clinical significance mainly relates to
DMEK patients in whom we have observed this sign in
the so-called stromal gutter between graft and host,
described by Ebru Cömert et al. [59] This phagocytic
ability of migrating endothelium should not be
mistaken for immunologic rejection in DMEK grafts.

OUTCOMES OF SUBSEQUENT
TRANSPLANTS AND CATARACT
EXTRACTION
A valid concern amongst corneal specialists regard-
ing this procedure is the impact on any subsequent
corneal graft that may be required. Early reports are
&
reassuring. In 2017 Rao et al. [51 ] reported out-
comes of DMEK in three patients after failed DSO
surgery. In the short term, two of three patients
FIGURE 8. Slit lamp photograph demonstrating endothelial achieved 20/20 vision with high cell counts. One
pigment accumulation within the Descemetorhexis zone. patient achieved 20/70 vision with postoperative
anterior chamber inflammation and cystoid macu-
lar edema that responded to topical therapy. In our
Endothelial pigment deposition own center, since 2014, three grafts have been per-
In several eyes following DSO, we have observed a slow formed for nonresponding DSO patients. One
uptake of pigment granules in the stripped area. These patient received a DSEK after failing rescue therapy
are seen as small, darkly pigmented spots that are not with Rho-kinase inhibitor – compounded Y-27632
&&
affected by application of topical steroid (Fig. 8). The [44 ]. A DSEK was chosen because of the presence of
ability of the human corneal endothelium to phago- Descemet’s fibrosis and local detachment. Vision
cytose circulating aqueous pigment has been docu- remains 20/25 at 3 years. Two patients failed to
mented [58]. Why this is seen in the central stripped respond to the procedure despite the use of topical
area post-DSO is unclear. There may be an increased ROCK inhibitor from day 1 postoperatively. Both
phagocytic ability in migrating cells with more filo- received a DMEK with vision of 20/20 and 20/25
podia, or this may simply be a consequence of restored (Fig. 9). More detailed analysis and

FIGURE 9. (a) Slit lamp photograph 12 weeks post Descemet’s stripping only with no formation of clear zone, with decision
made to proceed to Descemet’s membrane endothelial keratoplasty (DMEK). (b) Slit lamp photograph of same patient 8 weeks
post DMEK demonstrating clear cornea and restoration of vision to 20/25.

1040-8738 Copyright ß 2019 Wolters Kluwer Health, Inc. All rights reserved. www.co-ophthalmology.com 283

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.

Corneal and external disorders

2. Joyce NC. Cell cycle status in human corneal endothelium. Exp Eye Res
reporting of these outcomes will follow and are 2005; 81:629–638.
important to reassure our patients that the choice 3. Joyce NC. Proliferative capacity of the corneal endothelium. Prog Retin Eye
Res 2003; 22:359–389.
of a DSO does not compromise any future endothe- 4. He Z, Campolmi N, Gain P, et al. Revisited microanatomy of the corneal
lial keratoplasty. Combining this procedure or stag- endothelial periphery: new evidence for continuous centripetal migration of
endothelial cells in humans. Stem Cells 2012; 30:2523–2534.
ing this procedure with cataract surgery does not 5. Amann J, Holley GP, Lee SB, Edelhauser HF. Increased endothelial cell
& && &&
seem to affect outcomes thus far [20 ,36 ,46 ]. density in the paracentral and peripheral regions of the human cornea. Am J
Ophthalmol 2003; 135:584–590.
6. Tong CM, Melles GR. Where is endothelial keratoplasty going: from Desce-
met stripping (automated) endothelial keratoplasty to Descemet membrane
CONCLUSION endothelial keratoplasty to Descemet membrane endothelial transfer? Can J
Ophthalmol 2012; 47:197–200.
The transition from penetrating to lamellar kerato- 7. Parker J, Verdijk RM, Muller TM, et al. Histopathology of failed Descemet
plasty has become established amongst corneal sur- membrane endothelial transfer. Eye Contact Lens 2018; 44(Suppl.
1):S361–S364.
geons. We may now be taking our first steps beyond 8. Dirisamer M, Yeh RY, van Dijk K, et al. Recipient endothelium may relate to
lamellar grafting into cell-based therapies. DWEK/ corneal clearance in descemet membrane endothelial transfer. Am J Ophthal-
mol 2012; 154:290–296.e1.
DSO supplemented with topical Rho-kinase inhibitor 9. Dirisamer M, Ham L, Dapena I, et al. Descemet membrane endothelial
is likely an early step on that journey. What is cur- transfer: ‘free-floating’ donor Descemet implantation as a potential alternative
to ‘keratoplasty’. Cornea 2012; 31:194–197.
rently known is that the success of the procedure may 10. Birbal RS, Parker J, Dirisamer M, et al. Descemet membrane endothelial
relate to patient factors that remain poorly defined, & transfer: ultimate outcome. Cornea 2018; 37:141–144.
Ultimately all cases of Descemet’s membrane endothelial transfer required rescue
and surgical factors that are increasingly well defined, with endothelial graft, demonstrating unsatisfactory longevity of this procedure.
with avoidance of stromal contact paramount. The 11. Lam FC, Bruinsma M, Melles GR. Descemet membrane endothelial transfer.
Curr Opin Ophthalmol 2014; 25:353–357.
safety and efficacy of supplementary ROCK inhibitor 12. Okumura N, Okazaki Y, Inoue R, et al. Effect of the Rho-associated kinase
is the focus of current study with anecdotal reports of inhibitor eye drop (Ripasudil) on corneal endothelial wound healing. Invest
Ophthalmol Vis Sci 2016; 57:1284–1292.
success now under scrutiny and early trial results 13. Okumura N, Koizumi N, Ueno M, et al. Enhancement of corneal endothelium
encouraging. Both the cytoskeletal effects and the wound healing by Rho-associated kinase (ROCK) inhibitor eye drops. Br J
Ophthalmol 2011; 95:1006–1009.
mitotic effects of ROCK inhibition seem to be less- 14. Nakagawa H, Koizumi N, Okumura N, et al. Morphological changes of human
ened or even negated by cell–cell contact. This obser- corneal endothelial cells after Rho-associated kinase inhibitor eye drop
(Ripasudil) administration: a prospective open-label clinical study. PLoS
vation may be key to explaining why we have One 2015; 10:e0136802.
witnessed significant effects of these agents in the 15. Okumura N, Koizumi N, Kay EP, et al. The ROCK inhibitor eye drop accel-
erates corneal endothelium wound healing. Invest Ophthalmol Vis Sci 2013;
presence of an endothelial defect, but no resuscita- 54:2493–2502.
tion of low cell counts in situations of slow burning 16. 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
endothelial failure. Reports of the longevity of the 2015; 56:6067–6074.
procedure and health of subsequent grafts if required 17. Koizumi N, Okumura N, Ueno M, et al. Rho-associated kinase inhibitor eye
drop treatment as a possible medical treatment for Fuchs corneal dystrophy.
are necessary as the study of this technique continues. Cornea 2013; 32:1167–1170.
18. Koizumi N, Okumura N, Ueno M, Kinoshita S. New therapeutic modality for
Acknowledgements corneal endothelial disease using Rho-associated kinase inhibitor eye drops.
Cornea 2014; 33 Suppl. 11:S25–S31.
N.H. received the Peter Barry Fellowship from the Euro- 19. Garnock-Jones KP. Ripasudil: first global approval. Drugs 2014;
74:2211–2215.
pean Society for Cataract and Refractive Surgery. 20. Kaufman AR, Nose RM, Pineda R 2nd. Descemetorhexis without endothelial
& keratoplasty (DWEK): proposal for nomenclature standardization. Cornea
Financial support and sponsorship 2018; 37:e20–e21.
Proposed the term descemetorhexis stripping without endothelial keratoplasty as
Research into Fuchs’ Endothelial Dystrophy at Sydney standardized nomenclature for this procedure.
21. Moshirfar M, Ding Y, Shah TJ. A historical perspective on treatment of Fuchs’
Eye Hospital receives funding support from the Sydney endothelial dystrophy: we have come a long way. J Ophthalmic Vis Res 2018;
Eye Hospital Foundation. 13:339–343.
22. Paufique. Lamellar keratoplasty. In: Rycroft BW, editor. Corneal grafts.
G.M. is a consultant for Kowa Ind. manufacturers of England: Butterworth and Co., Ltd; 1955. pp. 132–133.
Glanatec (ripasudil). N.H. is a scientific consultant for 23. Braunstein RE, Airiani S, Chang MA, Odrich MG. Corneal edema resolution
after ‘descemetorhexis’. J Cataract Refract Surg 2003; 29:1436–1439.
Hoya Surgical and Carl Zeiss Meditec AG. 24. Patel DV, Phang KL, Grupcheva CN, et al. Surgical detachment of Desce-
met’s membrane and endothelium imaged over time by in vivo confocal
Conflicts of interest microscopy. Clin Exp Ophthalmol 2004; 32:539–542.
25. Zvi T, Nadav B, Itamar K, Tova L. Inadvertent descemetorhexis. J Cataract
There are no conflicts of interest. Refract Surg 2005; 31:234–235.
26. Pan JC, Au Eong KG. Spontaneous resolution of corneal oedema after
inadvertent ’descemetorhexis’ during cataract surgery. Clin Exp Ophthalmol
2006; 34:896–897.
REFERENCES AND RECOMMENDED 27. Agarwal A, Jacob S, Agarwal A, et al. Iatrogenic descemetorhexis as a complica-
tion of phacoemulsification. J Cataract Refract Surg 2006; 32:895–897.
READING 28. Srivastava VK, Singh A, Chowdhary R. Spontaneous resolution of corneal
Papers of particular interest, published within the annual period of review, have oedema after inadvertent descemetorhexis during cataract surgery. Med J
been highlighted as: Armed Forces India 2010; 66:177–179.
& of special interest 29. Choo SY, Zahidin AZ, Then KY. Re: Spontaneous corneal clearance despite
&& of outstanding interest
graft detachment in Descemet membrane endothelial keratoplasty. Am J
Ophthalmol 2010; 149:531.
1. Gordon SR. Cytological and immunocytochemical approaches to the study of 30. Shah RD, Randleman JB, Grossniklaus HE. Spontaneous corneal clearing
corneal endothelial wound repair. Prog Histochem Cytochem 1994; after Descemet’s stripping without endothelial replacement. Ophthalmology
28:1–64. 2012; 119:256–260.

284 www.co-ophthalmology.com Volume 30  Number 4  July 2019

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.

 Descemet’s stri pping without endothelial keratoplasty Garcerant et al.

31. Koenig SB. Long-term corneal clarity after spontaneous repair of an iatrogenic 46. Huang MJ, Kane S, Dhaliwal DK. Descemetorhexis without endothelial
descemetorhexis in a patient with Fuchs dystrophy. Cornea 2013; 32:886–888. && keratoplasty versus DMEK for treatment of fuchs endothelial corneal dystro-
32. Jullienne R, Manoli P, Tiffet T, et al. Corneal endothelium self-healing math- phy. Cornea 2018; 37:1479–1483.
ematical model after inadvertent descemetorhexis. J Cataract Refract Surg Compared visual outcomes of 12 DSO cases to 15 DMEK cases. Time to recover
2015; 41:2313–2318. vision was longer in DSO cohort with no significant difference in final visual acuity
33. Sanchez JC, Infantes Molina EJ, Cantos EA. Descemetorhexis without graft between groups.
placement: transplant of Descemet-free fragments. Ophthalmology 2017; 47. Macsai MS, Shiloach M. Use of topical Rho kinase inhibitors in the treatment
124:1620. && of Fuchs dystrophy after descemet stripping only. Cornea 2019;
34. Bleyen I, Saelens IE, van Dooren BT, van Rij G. Spontaneous corneal clearing 38:529–534.
after Descemet’s stripping. Ophthalmology 2013; 120:215. First prospective double arm cohort of DSO with and without ROCK inhibitor
35. Price FW Jr, Price MO. Spontaneous corneal clearance despite graft detach- (ripasudil) supplementation. Ripasudil cohort had faster time to recovery with
ment after descemet membrane endothelial keratoplasty. Am J Ophthalmol higher peripheral cell counts once healed.
2010; 149:173–174; author reply 174–175. 48. Soh YQ, Peh G, George BL, et al. Predicative factors for corneal endothelial
36. Borkar DS, Veldman P, Colby KA. Treatment of Fuchs endothelial dystrophy cell migration. Invest Ophthalmol Vis Sci 2016; 57:338–348.
&& by descemet stripping without endothelial keratoplasty. Cornea 2016; 49. Okumura N, Matsumoto D, Fukui Y, et al. Feasibility of cell-based
35:1267–1273. && therapy combined with descemetorhexis for treating Fuchs endothe-
A total of 10/13 cases cleared after 4-mm Descemet’s stripping only (DSO) lial corneal dystrophy in rabbit model. PLoS One 2018; 13:
combined with cataract surgery. Coined the terms fast, slow and nonresonders. e0191306.
37. Koenig SB. Planned descemetorhexis without endothelial keratoplasty in eyes In-vitro study deominstrating in-vitro repopulation of central cornea after desc-
with fuchs corneal endothelial dystrophy. Cornea 2015; 34:1149–1151. memtorhexis was possible. Repopulation was more successful if endothelial cells
38. Lee WB, Jacobs DS, Musch DC, et al. Descemet’s stripping endothelial but not Descemet’s membrane were removed.
keratoplasty: safety and outcomes: a report by the American Academy of 50. Sato T. Posterior incision of cornea; surgical treatment for conical cornea and
Ophthalmology. Ophthalmology 2009; 116:1818–1830. astigmatism. Am J Ophthalmol 1950; 33:943–948.
39. Melles GR, Ong TS, Ververs B, van der Wees J. Preliminary clinical results of 51. Rao R, Borkar DS, Colby KA, Veldman PB. Descemet membrane endothelial
Descemet membrane endothelial keratoplasty. Am J Ophthalmol 2008; & keratoplasty after failed Descemet stripping without endothelial keratoplasty.
145:222–227. Cornea 2017; 36:763–766.
40. Arbelaez JG, Price MO, Price FW Jr. Long-term follow-up and complications Good early outcomes of DMEK after failed DSO were demonstrated in three
of stripping descemet membrane without placement of graft in eyes with cases.
Fuchs endothelial dystrophy. Cornea 2014; 33:1295–1299. 52. Kaneko Y, Ohta M, Inoue T, et al. Effects of K-115 (Ripasudil), a novel ROCK
41. Moloney G, Chan UT, Hamilton A, et al. Descemetorhexis for Fuchs’ dystro- inhibitor, on trabecular meshwork and Schlemm’s canal endothelial cells. Sci
phy. Can J Ophthalmol 2015; 50:68–72. Rep 2016; 6:19640.
42. Malyugin BE, Izmaylova SB, Malyutina EA, et al. Clinical and functional results of 53. Okumura N, Fujii K, Kagami T, et al. Activation of the Rho/Rho kinase signaling
& one-step phaco surgery and central descemetorhexis for cataract and Fuchs pathway is involved in cell death of corneal endothelium. Invest Ophthalmol
primary endothelial corneal dystrophy. Vestn Oftalmol 2017; 133:16–22. Vis Sci 2016; 57:6843–6851.
Surgical study of DSO combined with cataract extraction, no supplementary Rho- 54. Okumura N, Koizumi N, Ueno M, et al. ROCK inhibitor converts corneal
associated kinase (ROCK) inhibitor was used. 63% cleared spontaneously, with endothelial cells into a phenotype capable of regenerating in vivo endothelial
the remainder requiring endothelial graft. tissue. Am J Pathol 2012; 181:268–277.
43. Iovieno A, Neri A, Soldani AM, et al. Descemetorhexis without graft placement 55. Okumura N, Okazaki Y, Inoue R, et al. Rho-associated kinase inhibitor eye
&& for the treatment of Fuchs endothelial dystrophy: preliminary results and drop (Ripasudil) transiently alters the morphology of corneal endothelial cells.
review of the literature. Cornea 2017; 36:637–641. Invest Ophthalmol Vis Sci 2015; 56:7560–7567.
Four of five cases cleared after 4-mm DSO. Concerns held for visual quality post 56. Breyer J, Samarin J, Rehm M, et al. Inhibition of Rho kinases increases
operatively due to irregular astigatism, procedure was not endorsed. directional motility of microvascular endothelial cells. Biochem Pharmacol
44. Moloney G, Petsoglou C, Ball M, et al. Descemetorhexis without grafting for 2012; 83:616–626.
&& fuchs endothelial dystrophy-supplementation with topical ripasudil. Cornea 57. Nakashima Y, Yoshitomi F, Oshika T. Clinical evaluation of cornea pseudo-
2017; 36:642–648. guttata. Br J Ophthalmol 2007; 91:22–25.
First description of DSO with supplementary ROCK inhibitor. Two failing cases of 58. Tripathi RC, Tripathi BJ. Human trabecular endothelium, corneal endothe-
DSO were rescued with grafts avoided after topical administration of ripasudil lium, keratocytes, and scleral fibroblasts in primary cell culture. A com-
drops. Clearance in 9/12 cases without ROCK inhibitor. parative study of growth characteristics, morphology, and phagocytic
45. Davies E, Jurkunas U, Pineda R 2nd. Predictive factors for corneal clearance after activity by light and scanning electron microscopy. Exp Eye Res 1982;
&& descemetorhexis without endothelial keratoplasty. Cornea 2018; 37:137–140. 35:611–624.
Analysis of patient and surgical factors in predicting clearance after DSO. No clear 59. Ebru Cömert CSS, Christopher S, Sáles MD, et al. Immunologic perspective
patient factor identified, all failing cases occurred with scoring rather than peeling on the denuded stromal ‘gutter’ after Descemet membrane endothelial
technique. keratoplasty. JCRS Online Case Rep 2015; 3:75–77.

1040-8738 Copyright ß 2019 Wolters Kluwer Health, Inc. All rights reserved. www.co-ophthalmology.com 285

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.