Oman Journal of Ophthalmology

: 2022  |  Volume : 15  |  Issue : 2  |  Page : 128--130

Ongoing debates in endothelial keratoplasty: DMEK is not the new DSAEK!

Haitham Al-Mahrouqi1, Charles N McGhee2,  
1 Department of Ophthalmology, University of Auckland, Auckland, New Zealand; Department of Ophthalmology, Al-Nahdha Hospital, Ministry of Health, Muscat, Oman
2 Department of Ophthalmology, University of Auckland, Auckland, New Zealand

Correspondence Address:
Dr. Charles N McGhee
Department of Ophthalmology, Faculty of Medical and Health Sciences, University of Auckland, Private Bag 92 019, Auckland
New Zealand

How to cite this article:
Al-Mahrouqi H, McGhee CN. Ongoing debates in endothelial keratoplasty: DMEK is not the new DSAEK!.Oman J Ophthalmol 2022;15:128-130

How to cite this URL:
Al-Mahrouqi H, McGhee CN. Ongoing debates in endothelial keratoplasty: DMEK is not the new DSAEK!. Oman J Ophthalmol [serial online] 2022 [cited 2023 Feb 2 ];15:128-130
Available from:

Full Text

Lamellar keratoplasty has been investigated for more than a century,[1] although it has only gained wide clinical acceptance over the last three decades. Compared to penetrating keratoplasty (PK), improved lamellar techniques that provide faster visual recovery, minimal astigmatism, better visual results, and lower rejection rates, are among the main drivers of increased adoption worldwide.[2] Notably, postoperative ametropia, including regular and irregular astigmatism, is still a major drawback of PK and deep anterior lamellar keratoplasty (DALK).[2] In comparison, since posterior lamellar keratoplasty, more commonly known generically as endothelial keratoplasty (EK), only involves changing the most posterior layers of the cornea, typically there is minimal refractive change, usually a minor hyperopic shift, in the overall refractive power of the eye.[3] Unlike PK, contemporary EK also eliminates long-term suture-related problems, potential wound dehiscence, and the demand for frequent postoperative review.

Since the first reports of EK in 1998 by Gerrit Melles,[4] the technique has advanced from insertion of a thick posterior disc through a large 9.0 mm sclero-corneal incision to inserting thinner grafts through small self-sealing corneal incisions. Descemet stripping endothelial keratoplasty (DSEK) or Descemet-stripping automated endothelial keratoplasty (DSAEK) - utilising a microkeratome, to create the donor lamellae, was subsequently described by Gerrit Melles in 2004.[5] The core technique involved stripping the host Descemet's membrane (DM) and the associated endothelium in a circular fashion (Descemetorhexis), leaving bare corneal stroma, followed by transplantation of a posterior lamellar corneal donor disc, containing endothelium, DM, and a thin layer of posterior stroma. Insertion was through a smaller 5 mm sclero-corneal tunnel with the corneal disc being held in place by apposition of the bare stromal surfaces and air bubble tamponade. Two years later (2006), Gerrit Melles described another novel technique of inserting a stripped roll of donor corneal tissue comprising only DM and endothelium, terming it Descemet's membrane endothelial keratoplasty (DMEK).[6] Currently, both DSAEK and DMEK are widely performed worldwide. However, the adoption of DSAEK was faster than DMEK for a number of reasons including: early uptake of tissue preparation by eye banks, ease of handling of the more substantial lamellar graft, aspects of recipients ocular anatomy, frequency of intra-operative complications, early visual recovery compared to PK, and survival of the graft. These factors are discussed in more detail hereafter.[7]

Tissue preparation of the corneal donor remains an important factor differentiating DMEK from DSAEK. With the introduction of microkeratomes, rather than manual preparation, DSAEK button preparation became simpler and more consistently repeatable, plus eye banks were quick to adopt preparation of precut tissue for surgeons. To produce thinner DSAEK buttons, the microkeratome can also be used to make a double pass, creating ultra-thin DSAEK (around 100 μm in thickness) for improved visual results.[8] Currently, DMEK relies on manual peeling techniques, which can be challenging and are more prone to failure. Typically, DMEK peeling is carried out by the surgeon in theater which consumes operating time. Nevertheless, eye banks around the world are increasingly preparing DMEK tissue and preloading it into injectable cartridges prior to surgery akin to DSAEK button preparation. Interestingly, young, diabetic, and pseudophakic donor eyes tend to be most difficult to use in DMEK as the DM scroll tends to tear out.[7] Thus, eye banks now typically use such eyes for other types of keratoplasties. The big bubble technique used in DALK can also be used to prepare an EK. However, pre-Descematic tissue will also be included in this technique, termed Pre-Descemet's endothelial keratoplasty (PDEK)PDEK.[9]

As the DMEK button/scroll is extremely thin (15–20 μ), compared to DSAEK and U-DSAEK buttons (70–135 μ),[7] handling of the graft tissue during insertion and positioning into the anterior chamber is more difficult with DMEK. Indeed, whereas DSAEK grafts are more rigid and therefore relatively easily oriented in the anterior chamber, DMEK grafts tend to scroll and usually require a higher degree of manipulation to achieve best orientation. In addition, DMEK grafts do not adhere well to remnants/rim of host DM, i.e., the respective Descemetorhexis diameter and donor centration is vital, and require a longer period of, and larger volume of, gas tamponade.[7] Thus, graft detachment is higher in DMEK especially in the early learning curve. It has been estimated that up to 20% of DMEK grafts require re-bubbling as compared to only 5% in DSAEK in Fuchs' endothelial corneal dystrophy (FECD) patients.[10]

The anatomy of the host eye may determine whether a DSAEK or DMEK is more suitable. Visibility is extremely important in DMEK to unscroll and correctly orientate the graft. Therefore, patients with reduced clarity due to significant corneal edema or corneal scarring may not be suitable for DMEK.[11] An irregular posterior corneal surface, as may occur in older penetrating keratoplasties or cornea trauma, may also make DMEK adherence more difficult.[12] Furthermore, as DMEK grafts tend to scroll tightly, they can be small enough to drop into the vitreous in eyes with aphakia or large iris defects. Finally, maintaining a good sized air/gas bubble and a pressurized anterior chamber in vitrectomized eyes and those with prior glaucoma filtration surgery can be challenging and hence such eyes may be more suitable for DSAEK.[12]

Ultimately, the postoperative results of successful DMEK are perhaps more appealing than DSAEK. Indeed, the transition from DSAEK to DMEK by many surgeons was lured by the faster postoperative recovery and better visual acuity. A study of fellow eyes, wherein one eye underwent DSAEK and the other eye underwent DMEK for FECD,[13] showed that patients with DMEK achieved slightly better Snellen mean best-corrected distance visual acuity (BCDVA) at 12 months (DMEK 20/23 vs. DSAEK 20/32). The patients also reported better quality of vision. Nevertheless, a disproportionate 33% of DMEK grafts compared to 7% of DSAEK grafts required re-bubbling in the postop period. Since this key study in 2011, techniques have been further refined for both EK techniques with improved outcomes and availability of longer-term results for comparison.

Over the last decade, perhaps unsurprisingly, it was noted that thinner DSAEK grafts (<131um) resulted in faster visual recovery and better BCDVA.[14] This is thought to be due to reduced stromal tissue and interface disparity which can degrade the quality of vision. This led to head-to-head randomized controlled trials (RCT) looking at DMEK versus thin DSAEK techniques. Chamberlain et al.[15] randomized 50 eyes with FECD to either DMEK or ultra-thin DSAEK (thickness 60–90 μm). The authors reported DMEK achieving better BCDVA with 1.5, 1.8 and 1.4 Snellen lines at 3-month, 6-month and 12-month as compared to ultra-thin DSAEK. Conversely, re-bubbling was significantly higher with DMEK (24%) compared to ultra-thin DSAEK (4%). There was no statistical difference in the rejection or failure rates. Similar results were replicated from another RCT using a mean DSAEK graft thickness of 100 um.[3] The hyperopic shift was 0.22D for DMEK and 0.58D for DSAEK. Overall, these results certainly suggest a useful vision advantage in the region of 1–1.5 Snellen lines for DMEK compared to DSAEK but at the cost of 1:4 compared to 1:20 re-bubbling rate postoperatively.

Survivability of transplants depends significantly on the initial surgical trauma, rejection episodes and endothelial attrition rate. A recent meta-analysis[10] comparing DMEK vs. DSAEK for FECD found that primary graft failure was 3.8% and 0.8% for DMEK and DSAEK respectively, although almost a five-fold difference this did not reach statistical significant. On the other hand, the risk of rejection was 5.5% and 8.2% for DMEK and DSAEK respectively, a statistically significant difference.[10] Yet, the absolute risk reduction is still relatively low. With regards to endothelial cell loss, there was not statistically significant difference between the two procedures.[10]

In conclusion, the debate in respect to the superiority of DSAEK or DMEK will still continue given the emergence of new clinical evidence and further surgical refinement in both techniques. Although Gerrit Melles reasonably predicted in 2006 that DMEK will eventually replace DSAEK,[16] DSAEK is still widely practiced. The relatively slower adoption of DMEK as a replacement for DSAEK is partly because of greater primary failure and technique difficulties in the learning curve.[17] Furthermore, the main benefit of improved visual outcome in DMEK is relatively modest (1–1.5 lines) moving from DSAEK to DMEK, and nowhere near to the transformational difference achieved in the early adoption of DSAEK in preference to PK. In addition, unlike EK versus PK the difference in rejection rate between DMEK and DSAEK is modest.[10] Finally, there is a trade off with further potential complications, including primary graft failure, due to difficult manipulation especially in anatomically abnormal eyes (e.g., aphakia, post vitrectomy, corneal scarring) and increased detachment rates.[12]

Ultimately, it is likely that both EK techniques will continue to survive side by side for the next decade or so, but DMEK is certainly not the new DSAEK, and each variant offers different advantages and disadvantages. We anticipate DMEK will become increasingly refined and popular, but each EK technique will still be applied to its most suitable indications, until laboratory engineered corneal tissue or stimulation of endothelial cell regeneration prevails in treating corneal endothelial disease.[18]


1Crawford AZ, Patel DV, McGhee CN. A brief history of corneal transplantation: From ancient to modern. Oman J Ophthalmol 2013;6:S12-7.
2Espandar L, Carlson AN. Lamellar keratoplasty: A literature review. J Ophthalmol 2013;2013:894319.
3Dunker SL, Dickman MM, Wisse RP, Nobacht S, Wijdh RH, Bartels MC, et al. Descemet membrane endothelial keratoplasty versus ultrathin descemet stripping automated endothelial keratoplasty: A multicenter randomized controlled clinical trial. Ophthalmology 2020;127:1152-9.
4Melles GR, Eggink FA, Lander F, Pels E, Rietveld FJ, Beekhuis WH, et al. A surgical technique for posterior lamellar keratoplasty. Cornea 1998;17:618-26.
5Melles GR, Wijdh RH, Nieuwendaal CP. A technique to excise the descemet membrane from a recipient cornea (descemetorhexis). Cornea 2004;23:286-8.
6Melles GR, Ong TS, Ververs B, van der Wees J. Descemet membrane endothelial keratoplasty (DMEK). Cornea 2006;25:987-90.
7Price MO, Feng MT, Price FW. Endothelial keratoplasty update 2020. Cornea 2021;40:541-7.
8Busin M, Madi S, Santorum P, Scorcia V, Beltz J. Ultrathin descemet's stripping automated endothelial keratoplasty with the microkeratome double-pass technique: Two-year outcomes. Ophthalmology 2013;120:1186-94.
9Agarwal A, Dua HS, Narang P, Kumar DA, Agarwal A, Jacob S, et al. Pre-Descemet's endothelial keratoplasty (PDEK). Br J Ophthalmol 2014;98:1181-5.
10Marques RE, Guerra PS, Sousa DC, Gonçalves AI, Quintas AM, Rodrigues W. DMEK versus DSAEK for Fuchs' endothelial dystrophy: A meta-analysis. Eur J Ophthalmol 2019;29:15-22.
11Trindade BL, Eliazar GC. Descemet membrane endothelial keratoplasty (DMEK): An update on safety, efficacy and patient selection. Clin Ophthalmol 2019;13:1549-57.
12Kemer ÖE, Karaca EE, Oellerich S, Melles G. Evolving techniques and indications of descemet membrane endothelial keratoplasty. Turk J Ophthalmol 2021;51:381-92.
13Guerra FP, Anshu A, Price MO, Price FW. Endothelial keratoplasty: Fellow eyes comparison of descemet stripping automated endothelial keratoplasty and descemet membrane endothelial keratoplasty. Cornea 2011;30:1382-6.
14Neff KD, Biber JM, Holland EJ. Comparison of central corneal graft thickness to visual acuity outcomes in endothelial keratoplasty. Cornea 2011;30:388-91.
15Chamberlain W, Lin CC, Austin A, Schubach N, Clover J, McLeod SD, et al. Descemet endothelial thickness comparison trial: A randomized trial comparing ultrathin descemet stripping automated endothelial keratoplasty with descemet membrane endothelial keratoplasty. Ophthalmology 2019;126:19-26.
16Melles GR. Posterior lamellar keratoplasty: DLEK to DSEK to DMEK. Cornea 2006;25:879-81.
17Williams KA, Keane MC. Outcomes of corneal transplantation in Australia, in an era of lamellar keratoplasty. Clin Exp Ophthalmol 2022 [Epub Ahead of print].
18McGhee C, Zhang J. Considering the evidence base in corneal transplantation: A penetrating, lamellar or cellular future? Clin Exp Ophthalmol 2022 [Epub Ahead of print].