Descemet Membrane Endothelial Keratoplasty in Complex Eyes

Abstract

Purpose of review

To review the current literature on Descemet membrane endothelial keratoplasty (DMEK) in complex eyes.

Recent findings

DMEK surgery has become a standardized procedure in Fuchs endothelial adystrophy and simple bullous keratopathy. But eyes with more complex disease present unique intraoperative and post-operative challenges to the DMEK surgeon. Poor visualization during surgery, complex anterior segment anatomy, altered anterior chamber dynamics, glaucoma shunts, and congenital or iatrogenic missing or altered iris and lens make DMEK surgery extremely difficult to accomplish.

Summary

DMEK is feasible in complex eyes including advanced bullous keratopathy, eyes with history of glaucoma or vitreoretinal surgery, previous penetrating keratoplasty, uveitis, pediatric and congenital anterior segment disorders. The tools and methods reported in the literature to accomplish DMEK in complex eyes vary widely with no particular consensus or standardization of techniques. The outcomes noted for some of these conditions demonstrate the difficulty of the surgery and the uncertainty of long-term graft survival in complex eyes. Both surgical standardization and randomized prospective data will better help elucidate DMEK’s role in the corneal rehabilitation of complex eyes.

Introduction

Over the past two decades corneal transplantation has been revolutionized with the inception of endothelial keratoplasty (EK). Selective replacement of the posterior layers of the cornea with EK began to exceed penetrating keratoplasty (PKP) surgeries in the US according to the Eye Bank Association of America (EBAA) Statistical Report in approximately 2011. [1]

Advantages of EK compared with PKP include quicker recovery period, improved visual outcome, smaller incisions, lower rates of rejection, and less compromise to globe integrity.[2]

Descemet membrane endothelial keratoplasty (DMEK) involves the selective replacement of the Descemet membrane and endothelial layers. Visual acuity and time of recovery of DMEK compared to Descemet stripping automated endothelial keratoplasty (DSAEK)) are probably better.[3–5] DMEK surgery has become more popular due to standardized techniques, more accessible due to eye bank involvement in tissue processing, and more cost-effective since its conception in early 2000s.[6]

DMEK surgery is most frequently chosen to address endothelial decompensation secondary to Fuchs endothelial dystrophy (FED) and bullous keratopathy (BK) in uncomplicated eyes.[1] However, DSAEK surgery may still be the preference of many surgeons in eyes with anatomical abnormalities or complex surgical history.[7] There are known challenges to performing DMEK surgery in complex eyes, but with reported benefits of faster recovery, reduced corneal higher order aberrations, and better visual acuity in Fuchs eyes,[3, 8] there is an interest broadening DMEK’s utility. A paucity of data exists in terms of efficacy and safety for DMEK in complex eyes. This report will highlight the current surgical and outcomes evidence in complex eyes that undergo DMEK surgery.

Bullous Keratopathy

BK is one of two common indications for performing endothelial corneal transplantation along with FED. Most cases of BK result from loss of endothelial cells secondary to an anterior segment surgery including routine or complex phacoemulsification surgeries, trabeculectomies, or glaucoma shunt implants. But any intraocular surgery and particularly anterior chamber surgeries can cause direct trauma to corneal endothelium and ultimately lead to BK. [9] Additionally, uveitis[10], toxic anterior segment syndrome (TASS)[11] and anterior chamber intraocular lenses (ACIOL’s)[12] can cause acute or chronic failure of endothelium. In the US in 2019, 4.6% of PKP’s and 17.3% of EK’s were done for corneal edema after cataract surgery.[1]

DMEK surgery for BK is being performed increasingly over DSAEK given its superior visual outcomes, lower graft rejection and failure rates, and decreased reliance on steroids.[7]. Visual acuity outcomes after DMEK in BK vs. FED, are comparable in long-term follow-up, but endothelial cell and graft survival may be inferior in BK. [13**]

BK may pose certain challenges for the DMEK surgeon. The corneal edema associated with BK is often advanced enough to prevent good visualization of the anterior chamber. Techniques available to help with visualization include superficial keratectomy of central edematous epithelium,[14] and intraoperative OCT to identify the scroll configuration of the graft.[15] Other techniques that may help visualization are utilization of a light pipe from vitrector equipment, external collimated light source, or re-staining the graft with trypan in the anterior chamber if visualization is lost during DMEK maneuvers.

Prior glaucoma surgery

An eye with a previous history of glaucoma surgery increases risk for endothelial decompensation. [16]. DMEK surgery in these eyes is frequently complicated by corneal edema, peripheral anterior synechiae, tube position, length and functioning, iridectomies, and blebs from trabeculectomies [17]. These changes after glaucoma intervention may lead to intraoperative difficulties during a DMEK surgery such as difficult graft unfolding, bleb rupture, or early post-operative escape of air/gas tamponade leading to increased risk of graft detachment and re-bubbling. More recent techniques such as goniotomy with Kahook dual blade may put the anterior chamber at risk for retrograde bleeding from the angle during graft insertion and unfolding. The Corneal Preservation Time Study suggested that DSAEK with previous glaucoma surgery eyes have a lower 3-year graft success rate. [18**] While early results of DMEK surgery in patients with glaucoma may be better than DSAEK [19], long-term outcomes are not impressive; mean survival time was 25±11 months after glaucoma drainage devices and 31.3±8.6 months after trabeculectomy in one recent retrospective study.[20] Another report showed 28% survival rate in DMEK compared to 33% in DSAEK at 4 years after EK surgery in eyes with history of trabeculectomies or glaucoma drainage devices.[21]

Techniques for performing DMEK surgery after glaucoma surgery also vary. Some have proposed longer acting gases such as sulfur hexafluoride (SF₆)[22] gas or perfluoropropane (C₃F₈) [23] instead of air. The management of the air or gas bubble may also present a challenge as a presence of a glaucoma drainage device or a trabeculectomy may accelerate the dissipation/displacement of the air or gas fill from the anterior chamber. Repositioning or trimming of the glaucoma drainage tubes may be required to prevent physical contact with the graft during and after surgery and to optimize graft attachment and survival. (Fig.1)

Figure 1.

Trimming of glaucoma shunt tube in DMEK surgery. Tube is carerully stretched and cut (A) to promote a shorter residual tube that is peripheral to graft but not retracted out of the anterior chamber(B).

Eyes with Previous Vitrectomy

An intact iris-lens diaphragm and the support of the vitreous are very helpful for chamber depth control and graft un-scrolling. Performing a DMEK surgery in post-vitrectomized eyes is often made difficult by the anterior chamber dynamics. The eye will soften during surgery and the chamber will spontaneously deepen as it equilibrates with the vitreous space. Chamber shallowing techniques popularized by Yoeruek et al.[24] have to be modified in avitric eyes. Additionally, some eyes with longer axial lengths, vitreous degeneration, or yag capsulotomy may behave with similar dynamics. Challenges similarly exists in eyes with scleral or iris fixated intraocular lenses (IOL’s) that are essentially unicameral. Intraoperative hypotony, bubble and even graft migration posteriorly are possible. [25] Finally, extensive DMEK graft manipulation in these challenging eyes may accelerate cell loss and earlier graft failure. Advanced techniques have been described for DMEK in avitric eyes.

Placement of a pars-plana infusion can offset the spontaneous deepening of the anterior chamber and intraoperative hypotony by providing posterior-to-anterior pressure gradient with surgeon-controlled or continuous infusion. [26] (Fig. 2) Alternatively an air bubble injected behind the IOL may provide a similar pressure gradient.

Figure 2.

Placement of a pars plana infusion during DMEK in a previously vitrectomized eye. A. Infusion is not running, promoting a soft eye with collapsed cornea but still deepened chamber. B and C. Foot-pedal intiated infusion restores corneal regularity, shallows chamber and allows controlled unscrolling of DMEK graft.

Compression techniques including digital pressure on the anterior scleral [24] or central cornea [27] may also shallow the chamber allowing flattening and un-scrolling of the graft. We have found that central corneal compression with a re-usable 128 D fundus lens (RESIGHT, Zeiss, Dublin, CA, USA) works to simultaneously visualize, compress and un-scroll a DMEK graft in eyes with soft, deep chambers. (W. Chamberlain, unpublished technique, Fig. 3).

Figure 3.

Corneal compression with a 128 diopter fundus lens during DMEK surgery. A. Visualization and shallowing of the chamber during DMEK injection (inset shows example of lens). B and C. Curvature and transparency of fundus lens allow visualization and simultaneous uniform anterior chamber compression to flatten and unscroll DMEK graft.

DMEK grafts can be configured in an endothelium in or tri-fold configuration. Jabbour et al. have reported on a pull through technique with tri-fold grafts which would theoretically place the graft in a correct orientation for bubbling with minimal manipulation. [28**] This method has potential to assist in eyes with soft and deepening anterior chamber dynamics. (Fig. 4)

Figure 4.

Endothelial-in configuraton of DMEK graft. A. Graft is folded on a custom platform into a tri-fold configuration with endothelium facing inward. B. Graft is pulled into eye with irrigating microforceps. C. Graft can be held in appropriate configuration against posterior cornea in preparation for introduction of bubble. (Courtesy of D. Srikumaran)

Other creative techniques have also been reported including a polymethyl methacrylate (PMMA) diaphragm to create an artificial double anterior chamber which supports unfolding [24] or a phakic anterior chamber IOL as a temporary pupil barrier in aphakic eyes.[29] Hayashi et al. trialed a ‘double-bubble’ technique (small bubble anterior to the graft and a larger bubble posterior to the graft after its insertion). Rolling the smaller bubble would assist with graft unfolding while larger bubble pushes the graft anteriorly toward the cornea in a deep chamber.[30]

Some post-vitrectomy eyes have ACIOL’s which pose significant risk for endothelial contact with the PMMA surface during graft manipulation as well as bubble displacement behind the ACIOL or iris. Droutsas et al. reported feasibility in 7 DMEK patients with indwelling ACIOLs. 2/7 grafts needed re-bubbling, 1/7 graft failed after 2^nd re-bubble, and 1 eye failed with postoperative endophthalmitis.[31]

While the visual acuity benefits of DMEK in post-vitrectomy eyes may be attractive, the current literature evaluating complications and longer-term outcomes should evoke caution. Yoeruek et al., in a case series of 20 eyes, reported 10% iatrogenic primary and 20% late graft failure, and 10% glaucoma exacerbation.[32] Mimouni et al. reported significantly higher rates of retinal detachments, cystoid macular edema and graft detachments in DMEK vs. DSEK at 1 year follow-up.[33]

Penetrating Keratoplasty (PKP) with Endothelial Decompensation

EK has become a popular method to rehabilitate a PKP with endothelial failure.[34] The same benefits of EK exist in a post PKP eye that had good corneal optics prior to endothelial decompensation, including lower risk of open-sky surgery and transplant rejection, more rapid visual recovery, and fewer refractive changes. Compared to DSAEK, a DMEK graft is thinner and theoretically better able to fit and appose against the posterior irregular cornea in a post-PKP eye, and may reduce optical degradation by eliminating the stroma-stroma interface.

A recent comparison of DMEK vs DSAEK after PKP demonstrated a 58% persistent graft detachment in the DMEK cohort leading to primary graft failure/early regraft.[35]. The range of re-bubbling’s in other series is 36–100%.[36–38]

While Descemet remnants and unevenness at the posterior graft host interface may drive these detachments,[38] there is no consensus on whether undersizing or oversizing the DMEK graft compared to the PKP size or omitting Descemet stripping may provide better outcomes. Graft sizing and centration are likely important as a larger graft may overlie the uneven graft host interface and promote detachments, while a slightly off-center smaller graft will have the same overlap problem. We frequently undersize DMEK grafts by 0.5 mm to fit entirely inside the PKP posterior surface. Sorkin et al. has proposed a femtosecond laser- created Descemetorhexis to reduce retained host Descemet tags that may promote graft detachments.[39].

Uveitis

The incidence of EK during the 10 years after diagnosis with uveitis is very low (1.1%). [40] However, the risk for immunological rejection and endothelial decompensation are likely higher and could affect graft survival.[10] The presence of posterior and anterior synechia, glaucoma shunts, trabeculectomies, weak zonules and intraoperative fibrin release are more common in this population and may complicate surgery. Hennein et al. recently reported that infectious uveitis (especially viral) was a strong predictor of graft failure across multiple forms of keratoplasty.[41*] Given the lower immunological rejection rates reported in DMEK compared to DSAEK and PKP[42], DMEK is an attractive but possibly more surgically challenging option in patients with uveitis.

Pediatric population

Few reports in the literature assess the outcomes of pediatric DMEK. The theoretical immunological and visual benefits of the DMEK surgery may be of great significance in this young population. Challenges of this surgery include: smaller anterior segment, increased propensity for globe collapse given lesser scleral rigidity, and the increased risk of inducing an early cataract.[43] Additionally, the thin Descemet membrane may be more difficult to peel than in adults due to the immature thin posterior non-banded layer. [44] The two most common indications for pediatric EK are congenital hereditary endothelial dystrophy (CHED) and posterior polymorphous corneal dystrophy (PPCD)[45] They may cause more profound corneal edema/opacification causing intraoperative visualization problems. Intraoperative OCT has been reported as a tool to assist in visualization.[46] Further difficulties arise in the post-operative period as pediatric population may mount a more profound inflammatory response. Post-operative supine positioning is also a significant challenge in the pediatric population.[43]

In one series of 5 eyes (age 6–15 years), 2 eyes where Descemet was not peeled required re-bubbling, 4 eyes had anatomical success with corneal clearing and one eye had a primary graft failure.[43]. In another series of 12 eyes from 8 patients (age 3–8 years) 6/12 omitted Descemet stripping; and 1 of these required postoperative successful re-bubbling and 0/12 had intraoperative complications using intraoperative OCT. [46] Ferguson et al. reported a successful case of DMEK surgery in a 4-year old, where intracameral tissue plasminogen activator was utilized to prevent fibrin formation.[47]

Phakic DMEK

DMEK surgery in phakic eyes is an appealing option in younger patients with residual range of accommodation. The refractive change in the cornea that occurs after DMEK may affect the IOL choice for cataract surgery. Staging cataract surgery after DMEK may allow for more predictable refractive outcomes due to known hyperopic and myopic shifts after DMEK. [48, 49].

The surgery in phakic patients is often simplified by the easy shallowing of the chamber due to space-occupying crystalline lens. However, cataract progression is a reported side-effect of performing EK surgery due to manipulation in the anterior segment, insertion of air or gas bubble, and prolonged post-operative treatment with topical steroids. Moshiri et al. recently reported a 40% rate of cataract surgery within 2 years after DMEK.[50] In contrast, a much larger retrospective series from Melles’ group demonstrated 17% cataract surgery at 10 years after DMEK with no graft detachments resulting from phacoemulsification.

Other complex anterior segment disease

Standardized DMEK techniques in non-complex eyes require normal iris-lens anatomy to achieve graft unfolding and air or gas tamponade.[51] Missing or maldevelopment of iris and lens and deviations in anterior segment geometry can have effects on graft crowding, unfolding, positioning, as well as air bubble stability. There are few studies that detail approaches to these complex eyes.

Santaella et al. evaluated the outcomes of DMEK in aphakic and aniridic eyes. Their cumulative graft survival probabilities at 12 and 24 months were a dismal 44% and 17%, respectively.[52] Given the attractive visual acuity outcomes of ultrathin DSAEK,[3, 53**] this may continue to be a preferred option in eyes where iris and or lens are absent. Optimization and standardization of pull through and endothelium-in scrolls may become a tool for predictable positioning of DMEK grafts in these complex eyes in the future.[28**]

Discussion/Conclusion

Surgeons who have overcome the steep learning curve of DMEK in eyes with FED are often interested in the utility of DMEK in eyes with more complex anatomy and pathology. Clearly, performing DMEK surgery in these eyes will require a different tool set and different learning curves. These barriers along with fewer available studies on complex eyes are likely to slow adoption rate when outcomes from DSAEK surgery in similar eyes are better understood.

Given that DMEK is a newer surgery, long-term success is not well quantified. The risks for short- and long-term complications in complex eyes is likely higher. Preliminary data on both complications and long-term success of DMEK in complex eyes discussed in this article should arouse caution for surgeons who are eager to adopt. A more careful assessment with prospective controlled trials would allow corneal surgeons to understand benefits and risks. The NIH sponsored DETECT trial (ClinicalTrials.gov Identifier: NCT05289661 and NCT05275972) is a 6–7 center randomized controlled patient masked, evaluator masked clinical trial. It will assess endothelial cell survival as a primary outcome in eyes with more complex pathology randomized to DMEK vs. ultrathin DSAEK. A number of secondary outcomes including visual acuity, graft survival, corneal higher order aberrations and vision related quality of life will also be assessed with enrollment scheduled to begin in 2022. Prospective data from well controlled trials, standardization of techniques, and exploration of emerging therapies for endothelial disorders will better define DMEK’s role in complex eye disease.

Table 1.

DMEK surgery in complex eyes: Summary of feasibility, challenges, techniques and expectations

Disease or condition	Difficulty of surgery*	Intraoperative challenges/risks		Potential novel techniques:		Expectation for long-term success
Bullous keratopathy	+	Limited view due to corneal edema		Superficial keratectomy Intraoperative OCT Vitrectomy light pipe illumination Staining with trypan blue		Equivalent visual acuity outcomes compared to Fuchs dystrophy [13] Potentially worse graft survival compared to Fuchs dystrophy [13]
Prior glaucoma surgery	++	Limited view due to corneal edema Peripheral anterior synechiae Tube positioning	Early air/gas escape Risk of bleb rupture	Tube repositioning or trimming Longer acting gases (SF6 or C3F8) Intraoperative OCT		Less that 30% survival time after 4 years [28] Mean survival time of ~25 months after glaucoma drainage devices and ~31 months after trabeculectomy [20]
Prior vitrectomy	+++	Suboptimal anterior chamber dynamics Intraoperative hypotony Bubble migration Graft unfolding difficulties		Pars-plana infusion Digital pressure or pressure with 128 D fundus lens ‘Double-bubble’ technique	Pull-through endo-in technique Longer acting gases (SF6 or C3F8)	Significant early/ late graft failure rate [32,33] 10% glaucoma exacerbation [32] Increased risk for retinal detachments and CME [33]
PKP with endothelial decompensation	++	Descemet unevenness at graft-host interface Limited view		Undersize grafts to PKP size Femtosecond-assisted descemetorhexis Longer acting gases (SF6 or C3F8)		High persistent graft detachment rate [35]
Uveitis	Limited data ++	Presence of anterior or posterior synechiae Intraoperative bleeding/ fibrin release		TPA to prevent fibrin formation		Higher risk for immunological rejection, endothelial decompensation and graft failure
Pediatrics	Limited data +++	Limited view Smaller anterior segment Lesser scleral/corneal rigidity	Risk of causing cataract Difficult stripping Post-operative positioning	Intraoperative OCT Longer acting gases (SF6 or C3F8)		Limited data exists
Phakic DMEK	+	Risk of causing/accelerating cataract formation		Usual ease of chamber shallowing		Limited data exists
Complex anterior segment disease	+++	Graft crowding Graft unfolding difficulties	Air bubble instability and migration Anterior chamber dynamics	Pull-through/endothelium-in technique Longer acting gases (SF6 or C3F8)		Poor graft survival rate beyond 2 years in aniridic/aphakic eyes [52]

^*Difficulty of surgery-- + (moderate), ++ (difficult), +++ (very difficult)

Key Points.

• Given the excellent outcomes of DMEK surgery in routine eyes, there is an interest in utilizing this procedure in eyes with more complex pathology and anatomy.

• DMEK in eyes with complex anterior segments is feasible but also more challenging for the cornea specialist.

• Many special techniques have been reported that can be utilized to accomplish DMEK in eyes with altered anterior segment anatomy, but no consensus or standardization of their use has been established.

• There is a lack of data on short and long-term outcomes of DMEK surgery in complex eyes which should elicit caution among surgeons and encourage development of more prospective studies and standardized techniques.