Corneal allograft rejection is an alloimmune response of the recipient against the donor corneal tissue. The pathology is classified as a delayed-type hypersensitivity reaction and is a complication that occurs in a certain proportion of cases after corneal transplantation. The incidence of rejection after penetrating keratoplasty (PKP) is approximately 10–30%, making it one of the main causes of corneal graft failure. Rejection, glaucoma, and infection are the three major complications of corneal transplantation.
Corneal transplantation is among the most successful types of organ transplantation. For first-time PKP in low-risk eyes, the 5-year graft survival rate reaches approximately 95% 1). This high success rate is supported by the immune privilege of the cornea.
The main factors that constitute corneal immune privilege are as follows:
Immune privilege is not absolute. In high-risk eyes with corneal neovascularization, the graft failure rate at 3 years postoperatively can exceed 35%. The most common cause of graft failure is irreversible immunological rejection, and the rate of progression from rejection to graft failure is reported to be approximately 49% 1).
“Graft rejection” refers to the recipient’s specific immune response against the donor cornea. In contrast, primary graft failure is due to defects in the donor tissue itself, surgical trauma, or improper preservation, and is defined as the graft never becoming clear within 8 weeks postoperatively. Primary graft failure is not immune-mediated and occurs in approximately 0.1% of PKP cases 3).
The diagnosis of rejection is only established for grafts that have maintained clarity for at least 2 weeks postoperatively. More than half of rejection episodes occur within the first year after surgery, with a peak between 6 months and 1 year. However, there are cases where the first rejection episode occurs more than 20 years after surgery.
Corneal transplantation is the most frequently performed tissue transplant worldwide. According to an international survey in 2012, PKP accounted for approximately 70% of all corneal transplants 1). In recent years, DSAEK and DMEK for endothelial diseases have rapidly become widespread, and DALK for keratoconus and post-keratitis scars has become a standard option, so the composition of surgical procedures has changed significantly 1). However, PKP remains essential for extensive corneal opacities and deformities, and it carries the highest risk of rejection among surgical techniques 1).
For first-time PKP, the incidence of rejection is reported to range from approximately 10% to 30%, with most episodes occurring between 6 months and 1 year postoperatively. The rate of progression from a rejection episode to graft failure is approximately 49%, and grafts that have experienced even one rejection episode have reduced long-term survival 1). Therefore, early detection of rejection and prompt initiation of treatment are decisive factors for graft survival.
More than half of episodes occur within the first year after surgery, especially between 6 months and 1 year. However, rejection can also occur long after surgery, so if symptoms such as redness, blurred vision, or decreased vision appear even years later, prompt medical attention is necessary. There are reports of rejection triggered by vaccination more than 20 years after surgery 10).
Criteria for rejection include any of the following: redness, photophobia, decreased vision, anterior chamber cells, keratic precipitates (KP), endothelial or epithelial rejection lines, subepithelial infiltrates, and localized graft edema 1). Localized KP on the graft is the most characteristic feature, and the absence of KP on the recipient cornea is an important distinguishing point from viral endotheliitis.
Rejection is classified into three types based on the layer affected: epithelial, stromal, and endothelial. Endothelial rejection has the greatest impact on graft prognosis, and delayed treatment leads to irreversible endothelial failure and vision loss.
Epithelial rejection
Stromal rejection
Endothelial rejection
The Khodadoust line is a linear corneal endothelial deposit characteristic of endothelial rejection. It represents the advancing front of rejection that progressively moves across the graft endothelial surface. In the area where the line has passed, endothelial cells are damaged, leading to stromal edema. When a Khodadoust line is observed, prompt and intensive steroid therapy should be initiated.
Corneal transplant procedures are broadly classified into penetrating keratoplasty (PKP), deep anterior lamellar keratoplasty (DALK), Descemet stripping automated endothelial keratoplasty (DSAEK), and Descemet membrane endothelial keratoplasty (DMEK). Due to differences in the amount of donor tissue transplanted and immunogenicity, rejection rates vary significantly among procedures.
Penetrating keratoplasty (PKP)
Rejection rate: approximately 4.9–28.9%1, 3)
Characteristics: Transplants all layers including epithelium, stroma, Descemet’s membrane, and endothelium, resulting in the highest antigen load.
Onset: Peaks at 6 months to 1 year postoperatively; more than half occur within the first year.
Deep anterior lamellar keratoplasty (DALK)
Rejection rate: 1–24%4)
Characteristics: Preserves the donor endothelium, so endothelial rejection does not occur in principle.
Challenges: Stromal rejection can still occur. Descemet’s membrane perforation and double anterior chamber are characteristic complications.
Descemet stripping automated endothelial keratoplasty (DSAEK)
Rejection rate: average 10% (range 0–45%)3)
Primary failure rate: Mean 5% (range 0–29%)
Features: The carrier is the posterior stroma approximately 50–100 μm. Most cases are mild, with only scattered fine KP or pigmentation on the posterior corneal surface.
Descemet Membrane Endothelial Keratoplasty (DMEK)
Rejection rate: Mean 1.9% (range 0–5.9%)3, 7)
Primary failure rate: 1.7%
Features: Only Descemet membrane and endothelial layer are transplanted, minimizing antigen load. Suture-related rejection is absent because no sutures are used.
Large cohort studies have shown that DMEK has a significantly lower risk of rejection compared to PKP and DSAEK3). However, a meta-analysis of 8 studies (376 eyes) comparing UT-DSAEK (ultrathin DSAEK, graft thickness <130 μm) and DMEK found no significant difference in rejection risk at 12 months postoperatively2). In the same meta-analysis, DMEK showed better logMAR corrected visual acuity at 12 months than UT-DSAEK (mean difference −0.06, 95% CI −0.10 to −0.02), but the risk of rebubbling (air reinjection) was significantly higher in the DMEK group (OR 2.76, 95% CI 1.46–5.22)2). A Dutch multicenter randomized controlled trial of 54 eyes also found that the DMEK group had a significantly higher rate of achieving 20/25 or better at 12 months than the DSAEK group (66% vs 33%, P=0.02), while there were no significant differences in endothelial cell density or refractive change11).
Endothelial rejection and viral endotheliitis have similar clinical presentations; the pattern of KP attachment is the most reliable distinguishing feature.
| Finding | Rejection | HSV/VZV endotheliitis | CMV endotheliitis |
|---|---|---|---|
| KP distribution | Confined to graft | Also outside graft | Also outside graft |
| KP color | White to gray-white | Brown | Brown to white |
| Characteristic findings | Khodadoust line | Arlt’s triangle | Coin lesion |
In rejection, graft-localized KP is the most characteristic feature, and this distinguishes it from viral endotheliitis, where KP is present on the recipient cornea. Since corneal endothelial deposits may also be donor-derived at the time of corneal transplantation, recording the distribution of KP during daily examinations is extremely useful for differentiation.
Cases in which PKP is prone to rejection are called high-risk eyes, and the following factors are listed.
In DMEK, endothelial rejection is rare but can be triggered during steroid tapering. A case of rejection has been reported after switching from betamethasone to fluorometholone at 15 months post-DMEK7). Additionally, cases with peripheral anterior synechiae have been shown to have a higher risk of rejection after DMEK7).
In solid organ transplantation, the effect of HLA matching in suppressing rejection is well established, but in corneal transplantation, research results are inconsistent 1). The Corneal Transplant Follow-up Study II (CTFS II) conducted in the UK is a large-scale clinical trial that prospectively investigated the impact of HLA class II (HLA-DR) matching in high-risk PKP 1). From 1998 to 2011, 1133 transplants were accumulated, and under the condition of ≤2 HLA class I mismatches, patients were stratified into 0, 1, or 2 HLA-DR mismatch groups and assigned using the cohort minimization method 1). To avoid errors from serological methods, DNA-based techniques (PCR-SSP/PCR-SSO) were used for donor and recipient tissue typing 1).
In CTFS II, no clear relationship was found between the number of HLA-DR mismatches and the incidence of rejection 1). As shown in rodent models, multiple different immune pathways are involved in corneal transplant rejection, and this redundancy of immune responses is thought to partly explain the inconsistent results of HLA matching studies 1). On the other hand, there is a consensus that HLA class I matching tends to be beneficial in high-risk transplants 1).
Corneal transplant rejection after COVID-19 vaccination has been reported with mRNA vaccines (BNT162b2), viral vector vaccines (ChAdOx1), and inactivated vaccines (Sinopharm).
There are more than 20 case reports involving multiple vaccines and surgical techniques, and the majority recovered with steroid treatment 9). Although causality is not established, it has been hypothesized that systemic immune activation from vaccination may trigger rejection through cross-reactivity or non-specific immune activation against the transplanted cornea 6, 9).
Rejection reactions occurring 1 to 3 weeks after vaccination have been reported with mRNA vaccines, viral vector vaccines, and inactivated vaccines. Most cases recover with steroid treatment 9). Patients with a history of corneal transplantation are advised to consult their physician about increasing preoperative steroid eye drops and postoperative self-monitoring (vision, redness, eye pain).
The diagnosis of rejection is primarily made by slit-lamp microscopy. The following findings are systematically checked.
The main conditions requiring differentiation from endothelial rejection are as follows.
The most important distinguishing feature is the distribution of KP. In rejection, KP is generally confined to the graft, whereas in HSV/VZV endotheliitis, KP also attaches to the recipient cornea outside the graft. CMV endotheliitis is characterized by coin lesion-like KP and chronic persistent intraocular pressure elevation. If definitive diagnosis is difficult, comprehensively evaluate anterior chamber PCR, serum antibody tests, and response to steroid therapy.
Treatment of rejection is fundamentally based on anti-inflammatory therapy with steroids. For epithelial and stromal types, steroid eye drops alone are often sufficient, but for endothelial type, prompt anti-inflammatory treatment is essential to protect endothelial cells. Immunosuppressive drugs are not effective alone for acute treatment because they take time to take effect, and they are used in combination with steroids.
Mild (epithelial/stromal type)
Severe (endothelial type, Khodadoust line positive)
With early treatment, more than 50% of acute rejection episodes recover, but delayed treatment can lead to irreversible endothelial cell loss and graft failure. Patient education is important: patients should be aware of postoperative symptoms (redness, blurred vision, eye pain, photophobia) and seek early medical attention if any changes occur.
For prevention of rejection after corneal transplantation, a two-tier protocol based on risk stratification is used.
Postoperative Management for Normal-Risk Eyes
Antibacterial eye drops: Cravit ophthalmic solution 1.5% (levofloxacin) 5 times daily → taper and discontinue
Steroid eye drops: Rinderon ophthalmic solution 0.01% (betamethasone) 5 times daily → switch to Flumetholon ophthalmic solution 0.1% (fluorometholone) 2-3 times daily
Adjuvant: Lindeta PF otic/ophthalmic solution 0.1% used concomitantly when epithelial damage is severe
Systemic: Flumarin for intravenous injection 1 g/day (flomoxef sodium) infused for several days starting on the day of surgery
Postoperative management for high-risk eyes
Steroid eye drops: Initiate as for normal-risk eyes and continue for at least 1 year
Systemic steroids: Rinderon injection 0.4% 2 mg once daily intravenously for 3 days from surgery day, then Rinderon tablets 0.5 mg 2 tablets once daily tapered over 2 weeks
Cyclosporine A: Neoral capsules 25 mg 3 mg/kg/day, maintain trough level 70-100 ng/mL
Tacrolimus (for CsA-resistant cases): Prograf (tacrolimus hydrate) 0.05-0.1 mg/kg/day, trough level 8-10 ng/mL until 2 months postoperatively, then 5-6 ng/mL
Cyclosporine A is used in high-risk cases such as those with corneal stromal vascular invasion involving 2 or more quadrants, regrafts, history of rejection, or allogeneic limbal transplantation. The dose is adjusted based on C2 level (blood concentration 2 hours after oral administration) or trough level, and continued for about 6 months postoperatively. Systemic side effects, especially renal function, are monitored as needed.
Tacrolimus is used as a switch drug for cases that develop rejection while on oral cyclosporine. The target trough level is 8-10 ng/mL until 2 months postoperatively, and 5-6 ng/mL thereafter. Topical 0.03% tacrolimus eye drops are also used to prevent rejection in high-risk corneal transplantation.
Cyclosporine A 1% eye drops are useful as an alternative to allow early tapering of steroids in cases with steroid-responsive ocular hypertension. In an 18-year-old PKP case with bilateral simultaneous rejection, remission was achieved with methylprednisolone pulse therapy, then switched to CsA 1% eye drops for successful maintenance5). CsA 1% eye drops enable early tapering of potent steroids and contribute to long-term graft survival5).
After remission, long-term continuation of steroid eye drops is expected to suppress recurrence3). Sutures should be removed promptly when loosening or breakage is detected. Exposed sutures can trigger both rejection and late infection, so they are a key point in follow-up. Since local inflammation from suture removal can also trigger rejection, steroid and antibiotic eye drop therapy should be temporarily intensified after suture removal.
Suture management is a key factor influencing long-term prognosis after corneal transplantation. Using fluorescein staining with slit-lamp microscopy makes it easier to detect loose or broken sutures and surrounding epithelial damage. In principle, all continuous sutures should be removed, and it is desirable to perform removal one year or more after surgery when it can be done safely. To achieve better visual acuity after penetrating keratoplasty, reducing astigmatism is essential, and suture adjustment should be performed repeatedly from the early postoperative period while evaluating with mire rings and topography. Hyperopic astigmatism of 5 diopters or more is a good indication for complete suture removal.
Even after successful acute treatment of rejection, abrupt discontinuation of steroids can trigger recurrence, so tapering should be done carefully over weeks to months. When starting with betamethasone 0.1%, first continue a maintenance dose of 4 times daily for several months to a year, then switch to a low-concentration steroid such as fluorometholone 0.1% and maintain at 1–2 times daily long-term. If intraocular pressure rises, consider changing to loteprednol or adding glaucoma eye drops (prostaglandin analogs, beta-blockers, etc.).
In high-risk PKP eyes, the standard is to continue topical betamethasone 0.1% four times daily for at least one year, then switch to a low-concentration steroid (e.g., fluorometholone) for long-term maintenance. Systemic immunosuppressants (Neoral, Prograf) are continued for about six months postoperatively, adjusted while monitoring renal function and blood trough levels. Early dose reduction can trigger rejection, so tapering should be done cautiously.
The rejection rate for PKP is approximately 10–30% (range 4.9–28.9% in the literature), whereas for DMEK it is significantly lower at an average of 1.9% (range 0–5.9%) 2, 3). This difference is mainly due to the amount of donor tissue and antigen load transplanted. PKP transplants epithelium and stroma containing dendritic cells, resulting in high antigenicity, and sutures also trigger immune responses. In contrast, DMEK transplants only Descemet’s membrane and endothelium, with minimal antigen load and no sutures, reducing risk. However, even with DMEK, about 6% of cases experience rejection after steroid discontinuation, highlighting the importance of long-term steroid continuation.
The cornea maintains physiological immune tolerance through anterior chamber-associated immune deviation (ACAID). In ACAID, antigen-presenting cells become tolerogenic in a TGF-β-dominant environment, suppressing delayed-type hypersensitivity and complement-fixing antibody production against donor antigens. However, when high-risk factors such as neovascularization, inflammation, or suture loosening are present, this immune privilege is easily disrupted.
The central mechanism of rejection is delayed-type hypersensitivity, and the main effector cells are CD4+ Th1 cells. Activated Th1 cells produce IFN-γ, which induces MHC class II antigen-presenting cells throughout the transplanted cornea, leading to self-amplifying progression of the cellular immune response8). Dendritic cells are abundant in the superficial stroma and limbus, presenting donor-derived antigens to the recipient’s regional lymph nodes to establish sensitization. Activated effector T cells infiltrate the transplanted cornea from limbal blood vessels and damage donor endothelial cells and stromal cells.
The involvement of antibody-mediated mechanisms has also gained attention in recent years. It has been suggested that anti-HLA antibodies may cause chronic endothelial cell damage via complement activation, which could contribute to late endothelial failure after long-term follow-up1). The concept of antibody-mediated rejection established in solid organ transplantation is now being applied to corneal transplantation1).
Differences in rejection rates among surgical techniques are primarily due to the amount and antigenicity of transplanted donor tissue3).
DSAEK donor lenticules are created using a microkeratome to produce a free cap of 300–350 μm thickness, and the remaining approximately 100 μm is used for surgery. The donor cornea is punched with an 8 mm trephine, inserted into the anterior chamber using specialized pull-through instruments (e.g., Busin glide, NS Endo-Inserter), and attached to the posterior corneal surface with air tamponade. In DMEK, the Descemet membrane and endothelial cell layer are stripped, and the graft stained with trypan blue is injected from the anterior chamber. Rebubbling (air reinjection) after DMEK may be required to repair graft detachment, and meta-analysis has shown it is significantly more frequent in the DMEK group than in the UT-DSAEK group (OR 2.76, 95% CI 1.46–5.22) 2).
Peripheral anterior synechiae (PAS) have been noted as a risk factor for rejection after DMEK. In a mouse corneal transplantation model, the group with PAS showed significantly increased rejection, suggesting that direct contact between the iris and donor endothelium via PAS induces cytotoxic T lymphocyte activity and promotes rejection 7). Clinically, there have been reports of rejection in cases with PAS after DMEK 7).
Corneal transplant rejection is primarily mediated by cellular immunity, but rodent studies have identified multiple distinct immune pathways leading to rejection 1). This redundancy of immune response is considered one reason for inconsistent results in HLA matching studies 1). The CTFS II used a large cohort of 1133 transplants and DNA-based high-precision tissue typing to examine the impact of HLA class II matching, providing a foundation for deeper immunological understanding in the field of corneal transplantation 1). In recent years, the role of anti-HLA antibodies and antibody-mediated rejection has also gained attention, potentially leading to elucidation of the mechanism of late endothelial failure 1).
COVID-19 vaccination elicits a systemic immune response, inducing SARS-CoV-2 neutralizing antibodies as well as antigen-specific CD8+ and Th1-type CD4+ T cell responses 6). It is presumed that this immune activation may trigger rejection via cross-reactivity or nonspecific immune activation against the transplanted cornea 6). For inactivated vaccines, the immunogenicity of the adjuvant (aluminum hydroxide) may also contribute 9). However, at the meta-analysis level, no increase in rejection after COVID-19 vaccination has been confirmed in solid organ transplantation, and a causal relationship in corneal transplantation has not been established at present.
Case accumulation of corneal graft rejection associated with COVID-19 vaccines is progressing worldwide, with at least 20 cases reported 9). Most cases are regrafts, occurring 1–2 weeks after vaccination, and the majority recover with steroid treatment 9). Prophylactic steroid increase before vaccination has been proposed, but no randomized controlled trials exist, and decisions must be made on a case-by-case basis 8, 9).
Regarding the clinical significance of HLA matching, the CTFS II has completed a large-scale prospective validation 1). Currently, no clear clinical benefit of HLA-DR matching in corneal transplantation has been demonstrated, but the role of anti-HLA antibodies and antibody-mediated rejection is becoming clearer, which may lead to elucidation of the mechanism of late endothelial failure and discovery of new therapeutic targets 1).
In comparisons between DMEK and UT-DSAEK, both the Sela 2023 meta-analysis 2) and the Dunker 2020 multicenter RCT 11) showed no significant difference in 12-month rejection rates, while the DMEK group achieved better corrected visual acuity. However, graft failure tended to be slightly higher in the DMEK group 2), and the risk of rebubbling (air reinjection) was also higher in the DMEK group (OR 2.76) 2). Surgical technique selection should be based on a comprehensive assessment of the patient’s individual ocular pathology, history, and facility experience 2, 11).
The following directions are attracting attention for the future.
