Corneal transplant: the whole cornea is no longer replaced, and that changes everything

The word “transplant” is frightening, and it is misleading: in most of the procedures performed today, the cornea is not replaced, only the one layer that is diseased. A fifteen-micron sheet for Fuchs’ dystrophy, a disc of stroma for advanced keratoconus. This shift, which began in the mid-2000s, has changed the three answers every patient is waiting for: rejection, the vision that comes back, and how long the graft will last.

UNDERSTANDING

What has changed: transplantation has become surgery on a single layer

Until the mid-2000s, there was only one operation: cut a full-thickness disc out of the cornea, remove it, and stitch in its place the same disc taken from a donor. This is penetrating keratoplasty, first performed successfully in 1905. It replaced all five layers, including the healthy ones, much as one would replace an entire windowpane for a flaw on its inner face.

Two innovations ended that monopoly. DSEK was born in 2004: only the recipient’s Descemet’s membrane is stripped away, to be replaced with a thin graft; its automated version, DSAEK, followed two years later. In 2006, DMEK took the logic to its conclusion and transplanted nothing but Descemet’s membrane and its endothelium, with no stroma at all. In parallel, deep anterior lamellar keratoplasty, or DALK, tackles the other side of the problem: replacing the stroma while leaving the patient’s endothelium in place.

The shift is documented. In the United States, between 2015 and 2020, the share of penetrating keratoplasty fell from 40% to 36.4% of corneal transplants, and that of DSAEK from 47% to 33.9%, while DMEK rose from 9.8% to 27.8%. The figures a relative remembers from a transplant done twenty years ago therefore probably do not apply to you.

ANATOMY

Why the diseased layer dictates the operation

The cornea is a transparent window a few hundred microns thick that provides most of the eye’s optical power. It has five layers: the epithelium on the surface, Bowman’s layer, the stroma, which accounts for most of the thickness, Descemet’s membrane, and the endothelium on the inner face. Each one goes wrong for different reasons, and each gives a different disease.

The endothelium is a single layer of cells that pump water out of the tissue and keep it transparent. In humans, these cells do not divide: their number can only fall. Below a certain threshold, the cornea becomes waterlogged and cloudy, and vision turns milky, more blurred on waking. That is the mechanism of Fuchs’ dystrophy, and of the decompensation that sometimes follows surgery inside the eye.

The stroma, for its part, poses a problem of shape and clarity: it warps in keratoconus and turns opaque after an infection or an injury, while the endothelium often remains intact. Replacing the endothelium would be pointless and costly, because every manipulation of it destroys cells that will not grow back. Hence the rule that structures modern surgery: touch only what is diseased.

INDICATIONS

Who needs a transplant, and when

Three situations dominate. The large global survey published in 2016, covering 184,576 transplants performed in 2012, ranks the indications: Fuchs’ dystrophy 39%, keratoconus 27%, after-effects of infectious keratitis 20%. The rest is divided between trauma, failed earlier grafts, and rarer dystrophies.

Keratoconus leads to a transplant only late in its course, and less and less often: rigid contact lenses, cross-linking, and intracorneal ring segments make it possible to postpone or avoid one in some patients. A transplant is considered when vision can no longer be corrected optically, or when the apex of the cone has become opaque. Hydrops, a sudden rupture of Descemet’s membrane, affects roughly 2.5% of people with keratoconus and frequently leaves a scar that makes surgery unavoidable.

Fuchs’ dystrophy shows itself as a morning fog that takes longer and longer to lift, then as constant impairment. It has one dreaded peculiarity: cataract surgery, even when perfectly performed, uses up endothelial cells and can tip an already fragile cornea into decompensation. Hence the endothelial cell count before the operation in these patients, and, sometimes, the discussion of combining the two procedures.

No universal numerical threshold triggers the operation. The decision rests on the real impact: what you can no longer do, the vision in the other eye, the state of the optic nerve and the retina, because a transplant restores the transparency of the cornea, not the function of what lies behind it.

ENDOTHELIUM

DMEK and DSAEK: grafting a sheet just a few microns thick

These two techniques treat the same problem, an endothelium that has stopped pumping, and differ in the thickness of the graft. The DMEK graft consists of nothing but Descemet’s membrane and its endothelium, ten to fifteen microns. The DSAEK graft carries a slice of posterior stroma with it: around two hundred microns, or about a hundred in the so-called ultrathin version. The eye is not laid open, and the graft is held in place by an air bubble, with no sutures on the visual axis.

The American Academy of Ophthalmology’s report on DMEK gives the most useful orders of magnitude: mean corrected visual acuity of 20/21 to 20/31 depending on the series, and 37.6% to 85% of eyes at 20/25 or better at six months. Endothelial cell loss reached 33% at six months, rejection affected 1.9% of eyes, and primary graft failure 1.7%.

The most frequent complication is not rejection: it is partial detachment of the graft, which required a further injection of air in 28.8% of patients. This procedure, known as rebubbling, takes a few minutes in the operating room; it is routine and does not signal failure. A meta-analysis found 33.7% of eyes needing rebubbling after DMEK versus 11.0% after ultrathin DSAEK.

In return, DMEK gives better acuity: 0.07 logMAR at one year versus 0.14 after ultrathin DSAEK, with no difference in endothelial cell density. A Cochrane review, which found no randomized trial on this comparison, concludes more cautiously that there is a visual advantage supported by low-certainty evidence, obtained at the cost of more detachments.

Endothelial keratoplasty is performed in centers that specialize in corneal disease; I perform it at the Clinique Sainte-Geneviève (Paris 14th). What is at stake in the consulting room comes earlier, and counts just as much: recognizing a cornea that is starting to decompensate, working out how much comes from the cornea and how much from the rest of the eye, and deciding on surgery at the right moment rather than the last one.

STROMA

DALK: replacing the stroma, keeping your own endothelium

DALK involves removing the diseased stroma down to Descemet’s membrane, without opening the inside of the eye, and replacing it with donor stroma. The patient’s endothelium stays in place, and the consequence is decisive: endothelial rejection, the main cause of failure after penetrating keratoplasty, can no longer occur. That is what makes DALK the technique of choice in advanced keratoconus, where the endothelium is generally healthy.

A London team published 357 eyes operated on for keratoconus: 75.9% reached 20/40 or better within the year, and 81% after suture removal, with 4.2% graft failure over a mean follow-up of about twenty-two months. Forty-two surgeons, thirty-one of them trainees, performed the operations: these are not the results of a single operator.

The price to pay is technical difficulty. In that series, Descemet’s membrane was perforated in 45.4% of cases, and 24.1% of operations were converted to penetrating keratoplasty on the table. That conversion is not without consequences: the risk of rejection was multiplied by 1.94 and the risk of secondary glaucoma by 4.0.

Compared with penetrating keratoplasty in keratoconus, a meta-analysis of thirty studies finds fewer cataracts, fewer rejections, and fewer episodes of raised eye pressure after DALK, with better preservation of the endothelial cell count; the difference in final corrected acuity did not reach statistical significance. So DALK does not give better vision: it lasts longer and causes fewer complications. Its more demanding learning curve explains its slow adoption.

FULL THICKNESS

Penetrating keratoplasty has not disappeared, and for good reason

It remains indispensable when the full thickness is affected: deep scars after severe keratitis, opacities extending through both stroma and endothelium, perforations, trauma, and above all repeat transplants. Between 2017 and 2020 in the United States, redoing a previous graft had become its leading indication. Roughly a third of corneal transplants in the United States are still done this way.

Its main constraint is not the operation itself, it is what comes afterward. The graft is held by sutures that stay in place for many months, sometimes more than a year, during which healing is slow and the junction between graft and recipient remains mechanically fragile for a long time.

Astigmatism is the most stubborn obstacle. In a series of 104 eyes, mean topographic astigmatism was still 3.4 diopters at the end of follow-up despite selective suture removal; 22% of eyes needed relaxing incisions and 17% needed fitting with rigid or scleral contact lenses. Optical correction after penetrating keratoplasty is a course of treatment in its own right.

REJECTION

Rejection: what the word actually means

Rejection is a reaction of the immune system against the donor’s cells, not a mechanical expulsion of the graft. Its critical target is the endothelium: each episode destroys cells that are not replaced, and it is the accumulation of these losses that eventually costs the graft its transparency.

How often it happens depends first on what was transplanted. A review of series with at least five years of follow-up reports 1.7 to 13% rejection after DALK, 5 to 11.4% after DSAEK, and 1.7 to 2.8% after DMEK, whereas rejection remains the leading cause of failure after penetrating keratoplasty. The less foreign tissue is transplanted, the less there is for the immune system to react against.

The condition of the recipient’s eye matters just as much. In a first-time recipient whose cornea is not vascularized, graft survival at two years exceeds 90%; in the presence of risk factors (a cornea invaded by blood vessels, chronic inflammation, previous grafts), it falls to between 35 and 70%. In a series of 201 penetrating keratoplasties for keratoconus, new blood vessels appearing after the operation multiplied the risk of a rejection episode by 2.5.

A rejection episode is not the same as losing the graft. In that same series, 22.9% of eyes had a rejection episode, and yet 98.5% of grafts were still functioning. The condition is speed: treated early with intensive topical corticosteroids, a rejection is usually reversed. Hence the signs to know by heart (red eye, pain, sensitivity to light, drop in vision) and the instruction that goes with them: be seen the same day.

That leaves the question of vaccines, which comes up every time. Twenty-three cases of rejection occurring after a vaccination have been described, which fed a legitimate concern. But an American case-control study of 601 patients who had experienced rejection and 1,803 matched controls found no association: adjusted odds ratio 1.17, confidence interval 0.91 to 1.50.

VISION

What vision to hope for, and how soon

The speed depends on the technique; the final level depends above all on what is working well or badly elsewhere in the eye. After DMEK, the eye clears within a few days to a few weeks, and the glasses prescription changes little. After penetrating keratoplasty, you have to wait for the sutures to come out, which means months, before knowing what the final vision will be.

As for the level reached, a series of 350 endothelial keratoplasties followed for fifteen years gives a telling trajectory: median corrected acuity of 20/60 before the operation, 20/40 between the third and twelfth month, then 20/30 from the second year to the fifteenth, other eye conditions included. This is not brand-new vision: it is a lasting restoration of function in patients who had become unable to read.

A nuance that is rarely spelled out: the Cochrane review comparing endothelial keratoplasty with penetrating keratoplasty in Fuchs’ dystrophy found no solid evidence of a difference in final corrected acuity. What lamellar grafting demonstrably improves is higher-order optical aberrations and the speed of recovery.

The residual optical obstacle is almost always irregular astigmatism, and it mainly concerns penetrating keratoplasty. Glasses and soft contact lenses are rarely enough in marked cases; scleral lenses are often the most effective solution. If optical correction fails, there remain relaxing incisions, refractive surgery, and implantation of a toric intraocular lens.

A study of patient-reported quality of life has the last word: median scores of 84.2 after DMEK, 77.8 after DALK, 76.1 after DSAEK, and 70.6 after penetrating keratoplasty; but at equal acuity, the type of surgery was no longer associated with the score. What patients judge is what they see.

LONGEVITY

How long a graft lasts

There is no single figure, and being told “ten years” with no further detail should raise suspicion: survival varies more with the indication than with the technique. The European corneal transplant registry, covering 12,913 procedures, finds graft survival of 89% at two years, ranging from 98% in keratoconus to 80% after trauma.

Over the long term, the series of 350 endothelial keratoplasties already cited offers the only fifteen-year projection available: 7% failure at five years, 16% at ten years, 23% at fifteen years. The main factor in late failure was pre-existing glaucoma, with the risk multiplied by 4.4 when it was treated with medication and by 24 when it had been operated on.

On the penetrating keratoplasty side, a Japanese series of 509 eyes followed for twelve years gives 60.4% clear grafts, an overall figure that means little on its own. Every graft placed for keratoconus or corneal dystrophy was still clear at twelve years, compared with 51.7% for bullous keratopathy. It is your indication, not the average, that tells you something.

Regrafting is possible, but each successive graft holds up less well. In the Japanese series, twelve-year survival was 65.4% for a first penetrating keratoplasty and 43.4% for a second. A Chinese cohort of 243 regrafts finds five-year survival of 53.5% after penetrating keratoplasty, 66.6% after DALK, and 69.8% after endothelial keratoplasty, with rejection as the main cause of failure.

THE GRAFT

Where the graft comes from, and what I cannot tell you

The graft comes from a deceased donor. A tissue bank retrieves it, checks for the absence of transmissible disease, and counts the endothelial cells before approving the tissue. Two storage methods coexist: cold storage, between 2 and 8 °C, which allows seven to fourteen days, and organ culture, between 28 and 37 °C, which allows four to seven weeks.

The shortage is global and documented. The survey published in 2016 counted 184,576 transplants performed in 2012 from 283,530 corneas procured by 742 eye banks, while 53% of the world’s population had no access to corneal transplantation at all: one cornea available for every seventy needed. Splitting a single cornea between several recipients, which lamellar techniques make possible, would save up to a quarter of the donated tissue.

On waiting times in France, I would rather not assert anything. The sources available to me are international, and the availability of grafts depends on a national organization of donation that none of them describes: dressing up a foreign figure as a French one would be misleading you. This question has an answer, but it belongs to the team that will be taking care of you. Ask it in consultation.

Finally, the choice between DMEK, DSAEK, DALK, and penetrating keratoplasty is not made on published averages: it is made on your endothelial cell count, on the depth of the damage as measured by imaging, on your history of eye surgery, and on the surgeon’s experience with the technique under consideration.

Frequently asked questions

Cachan practice · Tel. +33 1 45 47 08 11

Scientific sources

• Maghsoudlou P, Sood G, Gurnani B, Akhondi H. Cornea transplantation. StatPearls, 2024. PMID 30969512
• Moshirfar M, Thomson AC, Ronquillo Y. Corneal endothelial transplantation. StatPearls, 2023. PMID 32965936
• Xiao G, Tsou BC, Soiberman US, et al. Keratoplasty in the United States: trends and indications from 2015 to 2020. Cornea. 2023;42(11):1360-1364. PMID 36730420
• Gain P, Jullienne R, He Z, et al. Global survey of corneal transplantation and eye banking. JAMA Ophthalmol. 2016;134(2):167-173. PMID 26633035
• Hos D, Matthaei M, Bock F, et al. Immune reactions after modern lamellar (DALK, DSAEK, DMEK) versus conventional penetrating corneal transplantation. Prog Retin Eye Res. 2019;73:100768. PMID 31279005
• Yin J. Advances in corneal graft rejection. Curr Opin Ophthalmol. 2021;32(4):331-337. PMID 33989234
• Gurnani B, Czyz CN, Mahabadi N, Havens SJ. Corneal graft rejection. StatPearls, 2023. PMID 30085585
• Alio JL, Montesel A, El Sayyad F, et al. Corneal graft failure: an update. Br J Ophthalmol. 2021;105(8):1049-1058. PMID 32788325
• Armitage WJ, Goodchild C, Griffin MD, et al. High-risk corneal transplantation: recent developments and future possibilities. Transplantation. 2019;103(12):2468-2478. PMID 31765363
• Deng SX, Lee WB, Hammersmith KM, et al. Descemet membrane endothelial keratoplasty: safety and outcomes. A report by the American Academy of Ophthalmology. Ophthalmology. 2018;125(2):295-310. PMID 28923499
• Marques RE, Guerra PS, Sousa DC, et al. DMEK versus DSAEK for Fuchs’ endothelial dystrophy: a meta-analysis. Eur J Ophthalmol. 2019;29(1):15-22. PMID 29661044
• Stuart AJ, Romano V, Virgili G, Shortt AJ. Descemet’s membrane endothelial keratoplasty (DMEK) versus Descemet’s stripping automated endothelial keratoplasty (DSAEK) for corneal endothelial failure. Cochrane Database Syst Rev. 2018;6(6):CD012097. PMID 29940078
• Hurley DJ, Murtagh P, Guerin M. Ultrathin Descemet stripping automated endothelial keratoplasty (UT-DSAEK) versus Descemet membrane endothelial keratoplasty (DMEK): a systematic review and meta-analysis. Eye (Lond). 2023;37(14):3026-3032. PMID 36934158
• Moura-Coelho N, Papa-Vettorazzi R, Reyes A, et al. Ultrathin DSAEK versus DMEK: review of systematic reviews. Eur J Ophthalmol. 2024;34(4):913-923. PMID 37964555
• Nanavaty MA, Wang X, Shortt AJ. Endothelial keratoplasty versus penetrating keratoplasty for Fuchs endothelial dystrophy. Cochrane Database Syst Rev. 2014;2014(2):CD008420. PMID 24526345
• Gadhvi KA, Romano V, Fernández-Vega Cueto L, et al. Deep anterior lamellar keratoplasty for keratoconus: multisurgeon results. Am J Ophthalmol. 2019;201:54-62. PMID 30721690
• Shams M, Sharifi A, Akbari Z, et al. Penetrating keratoplasty versus deep anterior lamellar keratoplasty for keratoconus: a systematic review and meta-analysis. J Ophthalmic Vis Res. 2022;17(1):89-107. PMID 35194500
• Yu AC, Spena R, Pellegrini M, et al. Deep anterior lamellar keratoplasty: current status and future directions. Cornea. 2022;41(5):539-544. PMID 34759197
• Fuest M, Mehta JS. Strategies for deep anterior lamellar keratoplasty after hydrops in keratoconus. Eye Contact Lens. 2018;44(2):69-76. PMID 28328727
• Alió JL, Niazi S, Doroodgar F, et al. Main issues in penetrating keratoplasty. Taiwan J Ophthalmol. 2024;14(1):50-58. PMID 38654981
• Asena L, Altınörs DD. Visual rehabilitation after penetrating keratoplasty. Exp Clin Transplant. 2016;14(Suppl 3):130-134. PMID 27805532
• Bineshfar N, Tahvildari A, Feizi S. Management of post-keratoplasty ametropia. Ther Adv Ophthalmol. 2023;15:25158414231204717. PMID 37854948
• Alsubaie H, Alrubaian A, Ahmad K, Ahad MA. Risk factors for corneal graft rejection after penetrating keratoplasty for keratoconus. Int Ophthalmol. 2024;44(1):286. PMID 38935251
• Lee EH, Li JY. Immunization-associated corneal transplantation rejection: a review. Cornea. 2022;41(5):660-663. PMID 34629440
• Ku JH, Nanji AA, Tubert JE, et al. Risk of corneal graft rejection and vaccination: a matched case-control study from a United States integrated health care system. Am J Ophthalmol. 2024;263:133-140. PMID 38417569
• Dunker SL, Armitage WJ, Armitage M, et al. Outcomes of corneal transplantation in Europe: report by the European Cornea and Cell Transplantation Registry. J Cataract Refract Surg. 2021;47(6):780-785. PMID 33278237
• Price MO, Price FW Jr. Descemet stripping endothelial keratoplasty: fifteen-year outcomes. Cornea. 2023;42(4):449-455. PMID 35389930
• Ono T, Ishiyama S, Hayashidera T, et al. Twelve-year follow-up of penetrating keratoplasty. Jpn J Ophthalmol. 2017;61(2):131-136. PMID 27885526
• Zheng Q, Chen W, Li J, et al. Repeat corneal transplantation in Southern China: indications, surgical technique, outcomes, and risk factors for repeat keratoplasty failure. Indian J Ophthalmol. 2024;72(Suppl 3):S482-S487. PMID 38648456
• Dunbar GE, Titus M, Stein JD, et al. Patient-reported outcomes after corneal transplantation. Cornea. 2021;40(10):1316-1321. PMID 33758138
• Cipolla J, Jo M, Sabater AL. A comprehensive review of donor corneal preservation strategies. Cornea. 2024;43(11):1456-1465. PMID 39102308
• Gadhvi KA, Coco G, Pagano L, et al. Eye banking: one cornea for multiple recipients. Cornea. 2020;39(12):1599-1603. PMID 32947412
• Borgia A, Coco G, Airaldi M, et al. Role of direct supervision in the learning curve of Descemet membrane endothelial keratoplasty surgery. Cornea. 2024;43(1):52-58. PMID 37098113
• Melles GR, Wijdh RH, Nieuwendaal CP. A technique to excise the descemet membrane from a recipient cornea (descemetorhexis). Cornea. 2004;23(3):286-8. PMID 15084862
• Gorovoy MS. Descemet-stripping automated endothelial keratoplasty. Cornea. 2006;25(8):886-9. PMID 17102661

This article is intended for information and is not a substitute for a medical consultation. Indications, expected benefits, and risks depend on each individual situation. For any question about your vision, consult an ophthalmologist.