Corneal transplantation
in 2026
A complete state of the art: from historic full-thickness replacement to selective, regenerative and bioengineered corneal medicine. Indications, techniques, tissue banking, immunology, complications — and the shift redefining the specialty.
Summary figure — one layer, one surgery
Modern logic can be read on a cross-section: each stratum now has a dedicated technique.
Watch the surgical videos illustrating these techniques in the Videos section of the site:
Historical evolution & the selective paradigm
Keratoplasty is the oldest and most successful of organ transplants: the first successful human graft was performed by Eduard Zirm in 1905, in Olomouc. For nearly a century, penetrating keratoplasty (PK) — full-thickness replacement of the cornea by a sutured full-thickness button — was the reference procedure, whatever the layer actually affected.
The conceptual turning point, from the 1990s-2000s, was the move to selective lamellar surgery: because the cornea is a stratified tissue with clearly separated functions (epithelial barrier, transparent stromal framework, endothelial pump), ideally one replaces only the failing stratum. This principle gave rise to DALK anteriorly, endothelial keratoplasties posteriorly, and, more recently, approaches that avoid any graft altogether.
In 2026, the clinician no longer asks "should we graft?" but "which layer is failing, and what is the most tissue-sparing intervention that restores it?" The entire decision tree flows from this question.
Anatomy & physiology applied to grafting
Five to six strata, each with a distinct regenerative capacity and function, dictate the whole surgical strategy.
| Layer | Thickness / nature | Regeneration | Consequence for grafting |
|---|---|---|---|
| Epithelium | ~50 µm, stratified squamous | Yes via limbal stem cells | Not grafted in isolation; limbal deficiency must be treated before any keratoplasty. |
| Bowman | ~10 µm, acellular collagen | No | Heals as fibrosis (haze); landmark of superficial lamellar keratoplasty. |
| Stroma | ~500 µm, 90% of thickness | Limited (keratocytes) | Transparent framework; target of DALK, ectasias, CAIRS and stromal substitutes. |
| Dua / pre-Descemet | ~10-15 µm, condensed lamellae | No | Cleavage plane of the type-1 big bubble; its description (2013) refined DALK and PDEK. |
| Descemet | ~10 µm, endothelial basement membrane | Reforms (scaffold) | Removed with the endothelium in DMEK; removed alone in DSO. |
| Endothelium | Monolayer, ~2500-3000 cells/mm² in adults | Not in vivo | The Achilles' heel: its loss is irreversible, hence the weight of endothelial indications and the major appeal of cell therapy. |
The dual key: transparency & the endothelial pump
Corneal transparency relies on (1) avascularity, (2) the regular arrangement of stromal collagen lamellae, and (3) a state of deturgescence maintained by the endothelial pump (Na⁺/K⁺-ATPase, bicarbonate transporters). As the endothelium does not divide in vivo, it works as an exhaustible reserve: below ~500 cells/mm², decompensation (oedema, bullous keratopathy) becomes inevitable. All endothelial surgery aims to restore this pump — by grafting donor cells (DSAEK/DMEK), by stimulating migration of the remaining cells (DSO ± ROCK), or by injecting cultured cells.
Corneal immune privilege
The cornea is an immunologically privileged site, which explains the excellent prognosis of grafts in standard settings. Its mechanisms: absence of blood and lymphatic vessels, low MHC expression, Fas-ligand expression (apoptosis of effector T cells), immunomodulatory factors in the aqueous humour (TGF-β, α-MSH), and anterior chamber-associated immune deviation (ACAID). This privilege collapses in the setting of neovascularisation, repeat grafting or inflammation — defining "high risk" (see Ch. 06).
The Dua layer (described in 2013) is the cleavage plane of the type-1 big bubble in DALK, and provides a more robust bed for PDEK. Understanding where the air bubble "lifts off" is understanding why a DALK succeeds or is converted.
Epidemiology & indications in 2026
The great registry shift is now confirmed: according to the Eye Bank Association of America statistical report for 2024, DMEK has become the most frequently performed keratoplasty (18 256 domestic procedures), ahead of DSAEK (16 345) and penetrating keratoplasty (14 143). Endothelial keratoplasties (DSEK+DMEK) total 41 558 procedures, up 4.2%, and now represent the clear majority of activity in developed countries. Penetrating keratoplasty, in continuous decline since its 2005 peak and overtaken by endothelial surgery from 2012, retains a residual but irreplaceable role.
| Category | Typical conditions | Preferred technique in 2026 |
|---|---|---|
| Endothelial | Fuchs dystrophy (leading indication, ~35%), pseudophakic bullous keratopathy, posterior dystrophy | DMEK ➜ DSAEK/UT-DSAEK if complex eye; DSO ± ROCK or cell therapy in selected cases |
| Stromal / ectasia | Keratoconus, pellucid degeneration, scars, stromal dystrophies (TGFBI…), post-infectious opacities | DALK (healthy endothelium); CAIRS or rings to remodel; stromal substitutes in trials |
| Combined / total | Perforation, pan-corneal opacity, multiple failure, very irregular cornea | Penetrating keratoplasty (often femto-assisted) |
| Surface / limbus | Limbal deficiency (burn, aniridia, Stevens-Johnson) | CLET (Holoclar®), SLET, KLAL — before any stromal graft |
| "Impossible" cornea | Repeat failures, major dryness, keratinised surface | Keratoprosthesis (Boston KPro, CorNeat, OOKP) |
| Infectious / urgent | Uncontrolled abscess, fungal/amoebic keratitis, perforation | "Hot" tectonic/therapeutic PK |
The structuring weight of scarcity
Graft availability is highly unequal: classic estimates (Gain et al., 2016) suggest a single graft available for about 70 patients worldwide, with recovery rates below 20% in many resource-limited countries. This scarcity — and not surgical performance alone — is the main driver of 2026 innovations: cell therapy (one donor for dozens of recipients), animal-derived or synthetic bioengineered corneas, and bioprinting. France relies on a network of tissue banks overseen by the Agence de la biomédecine (see Ch. 05).
Market dynamics mirror clinical practice: strong growth of lamellar surgery, rise of artificial/bioengineered corneas as a response to scarcity, and the first regulatory steps for cell therapy. "Grafting" is becoming a continuum from donor tissue to cell-therapy products and bioengineered devices.
Taxonomy of techniques
Mapping the whole field before going into detail. Five major families, organised by the layer replaced or treated.
- Abbreviations
- PK penetrating keratoplasty · ALK anterior lamellar keratoplasty · DALK deep ALK · DSAEK Descemet stripping automated EK · DMEK Descemet membrane EK · PDEK pre-Descemet EK · DSO/DWEK Descemet stripping/without EK · CAIRS corneal allogenic intrastromal ring segments · KPro keratoprosthesis.