Cornea And Ocular Surface•

Outcomes after reversed corneal graft rejection: a report from the European VISICORT project

Abstract

Objective This study aims to describe the outcome of corneal grafts, both low risk and high risk, after successfully reversed immunological rejection.

Methods Datasets on reversed rejection episodes in penetrating and endothelial keratoplasties between 2014 and 2019 (n=876) were extracted from the Adverse Immune Signatures and their Prevention in Corneal Transplantation database, which contains the prospectively and consecutively collected corneal transplants from five European centres. Stratified by the preoperatively determined risk status for immunological rejection, the outcome parameters analysed included visual acuity, intraocular pressure, endothelial cell density and central corneal thickness before and after reversed rejection episodes.

Results Fourty-seven (52%) out of a total of 91 identified rejection episodes were successfully reversed and were available for analysis (23 penetrating and 24 endothelial keratoplasties). No statistically significant change was found for any of the parameters studied between the values before and the values 3 months after the rejection episode, irrespective of the preoperative risk status.

Conclusion The outcome of corneal grafts that survive immunological rejection may be clinically indistinguishable from the state before immunological rejection, irrespective of graft type and risk status. These findings support clinicians by providing information on prognosis after reversed rejection episodes and by giving patients realistic expectations regarding the outcome.

What is already known on this topic

  • Corneal transplants can be divided into high risk and low risk with regard to the risk of immunological rejection. In both risk groups, adequate management can reverse rejection and preserve graft function.

What this study adds

  • This study shows that visual acuity, intraocular pressure, endothelial cell density and central corneal thickness after successfully reversed rejection are not clinically relevantly different from the prerejection state, regardless of risk status or graft type.

How this study might affect research, practice or policy

  • Knowing that graft function can return to prerejection levels facilitates patient education and motivation with regard to both therapy to suppress rejection and long-term prophylaxis.

Introduction

Do we know enough about the fate of corneal grafts after reversed rejection episodes? Although this knowledge is relevant in clinical practice for the counselling and medical management of transplant recipients, it remains scarce1 and originates primarily from retrospective single-centre studies analysing diverse outcome parameters. VISICORT (Adverse Immune Signatures and their Prevention in Corneal Transplantation) is a European Union-funded multidisciplinary research project. The overall goal is to investigate systematic immune profiling of transplant recipients using a variety of biological samples. The VISICORT patient cohort consists of corneal transplant recipients at five European centres. The cohort also provides an opportunity to expand our knowledge of clinical outcomes after reversed graft rejection, in a multicentre, prospective study. We, therefore, set out to assess the outcomes after successfully reversed corneal graft rejection episodes in patients enrolled in the VISICORT project.

Materials and methods

Study design

Following the Strengthening the Reporting of Observational Studies in Epidemiology guidelines, we conducted a descriptive quantitative subanalysis of the prospectively collected multicentre VISICORT database. Ethical approval for the VISICORT project was obtained from each responsible ethics committee (online supplemental file 1). All participants gave written consent for the collection, processing and publication of their data. Otherwise, patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.

Setting

Starting in 2014, the VISICORT project has recruited adult patients undergoing corneal transplantation (including penetrating keratoplasties (PK) (n=241), endothelial keratoplasties (EK) (n=635) and anterior lamellar keratopasties (n=27)) at five European centres (Aarhus University Hospital, Denmark; Bristol Eye Hospital, UK; Charité – Universitätsmedizin Berlin, Germany; Nantes University Hospital, France and Royal Victoria Eye and Ear Hospital, Dublin, Ireland). Data collection was performed according to uniform criteria at predefined time points (preoperative, 6, 12, 24 and 36 months after transplantation).

Treatment regimen

There was no standardised treatment protocol. The patients were treated according to the usual guidelines of the respective hospital. However, the clinical leaders of each site agreed on the following before the study started: Normal postoperative treatment included administration of topical steroids at least four times daily for a period of 3 months, which was then tapered off over a period of months. Normal rejection treatment included hourly topical therapy with dexamethasone for 1–2 weeks, which was then tapered off. In addition, systemic steroids could be used if clinically indicated.

Participants

All participants included in this study had experienced an immunological rejection episode before February 2021 and belonged to either the PK or EK group. To rule out primary graft failure, patients who did not have at least one unremarkable postoperative follow-up visit before the rejection episode were excluded. To rule out irreversible rejection reactions, patients who did not regain a clear graft within 90 days of the rejection episode were also excluded.

Variables and criteria

The outcome variables measured were visual acuity, intraocular pressure (applanation or air-puff tonometry, depending on centre), endothelial cell density (specular microscopy depending on centre) and corneal thickness (Scheimpflug or Optical coherence tomography (OCT) imaging, depending on centre) between the last appointment before the rejection episode and a follow-up at least 3 months after the rejection episode. With regard to intraocular pressure and corneal thickness measurements, the same method was used for each patient during the observation period. Endothelial rejection episodes were defined as acute events with keratic precipitates on the corneal graft but not on the peripheral recipient cornea (either scattered or in the form of a Khodadoust line) along with an increase in central corneal thickness and oedema. Stromal rejection was defined as an acute event with opaque stroma and an increase in corneal thickness but without endothelial keratic precipitates. Epithelial rejection was defined as an acute event affecting the donor epithelium with a line located near engorged limbal vessels that migrated across the graft-recipient interface. The data collected included the indication for transplantation, ocular condition at the time of surgery, concurrent surgery, systemic or topical immunosuppressive treatment and postoperative clinical status. Transplants were classified as low risk or high risk, depending on the indication for transplantation and the presence of clinical risk factors (online supplemental file 2). Postoperative events and findings potentially promoting a rejection episode were also documented (loose or broken sutures, scheduled suture removal, neovascularisation at the graft margin, epithelial defects or ulcers, infectious keratitis, newly diagnosed intraocular pressure elevation and cataract surgery).

Bias

Potential selection bias due to the inclusion of multiple centres and observers was addressed by a common VISICORT database with predefined input fields and criteria. Missing data were mainly due to staff shortages or postponed follow-up visits in the context of the COVID-19 pandemic. To minimise information bias, patients with missing follow-up data were excluded from the analysis. Overall, one-third of all patients with rejection had to be excluded. The details of the exclusion process are summarised in figure 1.

Figure 1
Request permissionrequest permission icon
Figure 1

Flow chart of participants. EK, endothelial keratoplasties; DMEK, Descemet membranes EK, DSAEK, Descemet stripping automated EK; PK, penetrating keratoplasties; VISICORT, Adverse Immune Signatures and their Prevention in Corneal Transplantation.

Statistical analysis

Dichotomous variables were summarised as ratios and percentages. Interval scaled variables were described as means and SD. Changes in visual acuity were assessed using bar graphs. Visual improvement was defined as a ≥2 lines (Snellen) difference between follow-up visits. Fisher’s exact tests were performed to compare group frequencies at >3 months after rejection with those before rejection as well as to compare the rates of successful reversion stratified by low-risk and high-risk procedures, and by PKs and EKs. A two-sided Mann-Whitney U test for non-parametric variables was used to compare means between low-risk and high-risk PK, as well as correspondingly between low-risk and high-risk EK. For within-group differences, t-tests were conducted comparing means at different time points. Kaplan-Meier curves were plotted to visualise graft survivals and a log-rank test was performed to compare the graft survival curves. A p<0.05 was considered statistically significant. Plots and statistical analyses were made using GraphPad Prism (V.10.1.2) statistical software package.

Results

Participants

Overall, 241 PK, 387 Descemet stripping automated EK (DSAEK) and 248 Descemet membranes EK (DMEK) were performed. In these 876 grafts, 91 rejection episodes were documented. Details about the selection process are visualised in figure 1. After applying the exclusion criteria, 47 of 91 (52%) reversed rejections with a completed follow-up before and at least 3 months after rejection were included in the analysis. The included grafts consisted of 23 (49%) PKs, 15 (32%) DSAEKs and 9 (19%) DMEKs. Due to the low number of DMEKs and DSAEKs, these were analysed together as EK 24 (51%).

Overall, the rate of successful reversion was significantly higher in the low risk group (n=41) compared with the high-risk procedures (n=33) (p≤0.0001, Fisher’s exact test) while no statistically significant difference was observed between PKs and EKs (p≥0.9999, Fisher’s exact test) (online supplemental file 3).

Demographics and baseline characteristics

Patient demographics, the indications for surgery and the clinical risk factors used to classify the transplants preoperatively as low-risk or high-risk transplants are summarised in online supplemental files 2 and 4, respectively. All EKs were elective procedures and thus classified as low risk. In contrast, 3 of 7 (43%) PKs were emergency procedures (including 1 regraft), and therefore, categorised as high risk. None of the patients had topical or systemic immunosuppressive treatment before transplantation.

Surgical details and events after transplantation

Surgical details and postoperative events occurring between transplantation and the rejection episode are summarised in online supplemental file 5.

Time intervals and immunosuppressive treatment at the time of rejection

The time interval between surgery, rejection episodes and follow-ups are visualised in online supplemental file 6. Across all reversed grafts, most rejections were observed in the second year after transplantation (n=22, 47%). The graft survival of PKs and EKs as well as of low-risk and high-risk grafts is visualised in Kaplan-Meier curves (figure 2A). There was no significant difference in graft survival rates over the time period between transplantation and rejection between PKs and EKs (p=0.6857, log-rank test) as well as low-risk and high-risk grafts (p=0.2770, log-rank test). The median time between transplantation and rejection was significant longer in the group of successfully reversed grafts (n=47) than in the group of failed grafts (unsuccessfully reversed grafts, n=27) (p≤0.0001, Mann-Withney test, figure 2B).

Figure 2
Request permissionrequest permission icon
Figure 2

Graft survival. (A) Kaplan-Meier curves showing the comparison of graft survival depending on the type of graft (reversed penetrating keratoplasties (PK)/endothelial keratoplasties (EK)) (right) and on the risk status (low-risk (LR)/high-risk (HR) grafts) (left). No significant difference was found across groups (p=0.6857, PK vs EK, log-rank test; p=0.2770, LR vs HR, log-rank test). (B) Kaplan-Meier curves showing the comparison of graft survival between successfully reversed grafts (n=47) and non-reversed grafts (n=27). The survival curves are significant different (p=0.004, log-rank test). Further, the median time to rejection is significantly shorter in the group of non-reversed grafts (n=27) compared with the successfully reversed grafts (p≤0.0001, Mann-Whitney U test).

42 of the 47 (89%) reversed transplants completed the 36 months follow-up.

At the time of rejection, about a quarter of each group was taking topical steroid eye-drops while none of the transplant patients were on systemic immunosuppression.

Outcome parameters

The course of visual acuity, intraocular pressure, endothelial cell density and central corneal thickness between the last appointment before and a follow-up at least 3 months after reversed rejection are summarised in table 1.

Table 1
|
Summary of outcome parameters

Overall, the change in visual acuity was not significant for all groups (table 1). Figure 3 shows the line changes in best-corrected visual acuity (BCVA) from before rejection to the follow-up after rejection for all transplants. Three or more months after rejection, 11 of 15 (73%) eyes in the PK low-risk group and 6 of 7 (86%) eyes in the PK high-risk group showed no change or gained at least one line in BCVA, while 4 of 15 27%) and 1 of 7 (14%) eyes lost 2 or more lines. With regard to EK: in the low-risk group, 13 of 19 (68%) eyes showed no change or gained at least one line in BCVA, while 2 of 19 (10%) showed gains of 2 or more lines; in the high-risk group, 2 of 4 (50%) eyes showed no change or gained at least one line in BCVA, while the other half showed a loss of more than 2 lines.

Figure 3
Request permissionrequest permission icon
Figure 3

Bar graphs showing changes in best-corrected visual acuity (BCVA). EK, endothelial keratoplasties; PK, penetrating keratoplasties.

With regard to the intraocular pressure, no statistically significant changes were observed in all four groups. Three or more months after rejection, all four groups showed mean normotensive intraocular pressure. Only one eye in the PK high-risk group and one eye in the EK high-risk group had elevated intraocular pressure >21 mm Hg, of which only the latter was related to topical steroid treatment.

Corneal thickness measurements were missing in 2 (13%) in the PK low-risk group and 3 (43%) in the PK high-risk group. For EKs, measurements were missing in 4 (20%) in the low-risk group and in 1 (25%) in the high-risk group. The remaining low-risk and high-risk transplants showed no significant change in central corneal thickness measurement between before and a follow-up 3 or more months after rejection.

The number of missing measurements was higher for endothelial cell density measurements: 13 (81%) eyes in the PK low-risk group and 5 (71%) eyes in the PK high-risk group had no endothelial cell density measurement, while both groups of EK lacked a measurement in about half of the cases 9 (45%) eyes in the low-risk group, 2 (50%) eyes in the high-risk group.

Immunosuppressive treatment 3 or more months after rejection is summarised in table 2. During the observation period, none of the PKs failed after rejection while 2 (10%) low-risk EKs underwent late onset failure after rejection.

Table 2
|
Immunosuppressive treatment 3 or more months after rejection: topical and systemic steroid treatment ≥3 months after acute graft rejection

Discussion

The present report assessed the outcomes after successfully reversed corneal graft rejection episodes in patients enrolled in the VISICORT project. It shows that grafts surviving immunological rejection may be clinically indistinguishable from the state before the rejection episode. Evidence of this finding is that clinical characteristics such as visual acuity and intraocular pressure as well as instrument-based measurements such as central corneal thickness and endothelial cell density did not change to a statistically significant extent after the rejection episode, although it is important to note the high proportion of missing data in some places (table 1). This result seems to be largely independent of the type of transplantation (PK/EK) and the risk classification (low risk/high risk). The fact that no rejection episodes were observed in the anterior lamellar keratoplasty group is in line with the generally lower rejection rate for this transplant group.2

The first outcome parameter on which the present work focused was visual acuity. Not only is improvement of vision the most common reason for transplantation,2 it is also the most important parameter for patients to judge the success of the procedure, besides pain relief and preservation of the eye. The third importance of visual acuity is that its decrease is the most common symptom of a rejection episode, followed by pain, redness and photophobia.3 4 Despite the importance of visual acuity in the context of corneal transplantation, most outcome analyses focus on the risks and mechanisms of immunological rejection, and data on visual acuity development are correspondingly sparse.1 Retrospectively collected data on the development of visual acuity after successfully reversed rejection episodes in PK come from two German studies and indicate that prerejection baseline values can be regained approximately 6 weeks after the episode.3 5 The data collected in the present prospective analysis for PK are in agreement with these results and even formally show a slight increase in the mean values of visual acuity, both for low-risk and high-risk situations. The present analysis also found a similar trend in mean visual acuity in low-risk EK, whereas the formal decrease in mean visual acuity in high-risk EK does not allow generalisation due to the currently still small case number of n=4. One reason for the widespread absence of visual acuity as an outcome measure in the literature may be that visual acuity is influenced by comorbidities and may, therefore, be a source of bias.4 This report, therefore, presents the change in visual acuity not only with means and SD but also proposes an additional presentation in the form of bar graphs (figure 3). These show changes in BCVA as percentages of lines gained or lost between baseline and follow-up and are, therefore, independent of the level of baseline visual acuity (online supplemental file 4). With the exception of the numerically under-represented high-risk EK, the bar graphs confirmed the picture that the majority of patients see as well after reversed graft rejection as before the rejection episode.

The second outcome parameter on which the present work focused was intraocular pressure. The associations between corneal transplants and intraocular pressure are multifaceted. First, disease patterns such as intraocular inflammation or, for example, viral infection can lead to both corneal decompensation and pathologically elevated intraocular pressure.6 Second, glaucoma leads to an increased rate of endothelial cell loss, even in the absence of transplantation,7 and is among the most common causes of graft failure, along with immunological rejection and infection.8 9 Third, increased intraocular pressure can be induced by corneal transplantation itself, mechanisms include the postoperative inflammatory response, the use of viscoelastic substances, distortion of the chamber angle with collapse of the trabecular meshwork and the formation of peripheral anterior synechiae.10 Fourth, topically applied steroids can lead to an increase in intraocular pressure in 10%–35% of patients. Topical steroids are necessary for therapy of the postoperative inflammatory reaction, low dose for prophylaxis of an immunological rejection and high dose for therapy of a rejection episode.8 However, despite a large literature on the risk factors and incidence of glaucoma after corneal transplantation, only a few, partly contradictory and retrospectively collected data on pressure development after immunological rejection exist. In an English study on 140 rejection episodes, 16 eyes (11.4%) of eyes were found to have ocular hypertension at the time of rejection, and 27 eyes (19.3%) of eyes required antiglaucoma medication 6 weeks later. The number of irreversible rejection episodes was reported as 6, suggesting that a substantial number of reversed rejection episodes also suffered from ocular hypertension in the follow-up examinations.7 A German study on 20 rejection episodes could not establish a direct relationship between immunological rejection and elevation of intraocular pressure.3 In an Australian study on 235 rejection episodes, mean intraocular pressure values ranged between 15 and 16 mm Hg both before and at the time of immunological rejection, regardless of whether rejection was reversed or not.11 This is in agreement with the findings of this study, which showed normotensive mean pressure values before and after the rejection episode. For the management of transplant patients, this means that a successfully reversed rejection episode has a good chance of remaining at a normal pressure level despite circulating inflammatory cells and intensified steroid therapy.

The third and fourth parameters focused on in the present work were endothelial cell density and central corneal thickness. In the context of corneal grafts, endothelial cell density is important for two reasons: first, the gradual decrease in endothelial cell number over time is significantly faster in transplanted corneas than in healthy, non-transplanted corneas.12 Second, of all corneal layers, the endothelium is by far the most frequent target of immunological rejection reactions,13 as confirmed in the present report (online supplemental file 6). The literature on endothelial cell density after reversed rejection episodes is based on retrospective studies and does not provide a consistent picture. In addition to studies showing a significant reduction in endothelial cell density,14 15 there are those finding no reduction16 and those in whose cohorts both stable and reduced cell numbers were found.17 The results of the present report fit into this picture in that there was a slight formal decrease in endothelial cell number for all four subgroups studied (PK/EK; low risk/high risk), although it must be noted that the underlying database is narrow. Much more important, however, is the finding that all mean values remained within a range that still allowed normal functioning of the endothelium. The meaning of these findings for the treating ophthalmologist is that a reversed rejection episode does not necessarily result in a dramatic decrease in endothelial cell density. Accordingly, the available data formally showed an increase in central corneal thickness in all groups studied. This is in agreement with the limited literature that observed either a return to prerejection values18 or a discrete but permanent increase in corneal thickness of up to 10% after reversed graft rejection.5

Two findings that seem counterintuitive at first glance and which affect all four outcome parameters examined deserve to be discussed in more detail. First, the preoperative risk stratification with regard to a possible immunological rejection (low risk/high risk) after transplantation does not seem to have any influence on the outcome parameters after a successfully reversed rejection. This finding is in line with the results of two earlier studies19 20 that investigated the influence of the by far most important risk factor for an immunological rejection, host bed vascularity.21 Although blood and lymphatic vessels allow bidirectional transport of antigens and effector cells and thus represent a high-risk situation for immunological rejection, the mere initiation of a rejection within this framework does not necessarily also increase the risk for a functionally worse outcome.20 We speculate that this is due to the massive increase in topical steroid therapy that is commonly used to treat rejection episodes: there may be a critical threshold of immunosuppression below which risk factors can become effective and above which they can be neutralised.

Second, a successfully reversed rejection episode even seems to improve some individual outcome parameters when compared with the baseline values before rejection. In terms of mean visual acuity values, there was an increase in visual performance (not statistically significant) with PK, although an immunological attack on donor tissue is unlikely to result in functional improvement. However, the improved vision over time may also be due to progressive smoothing out of the cornea with improving astigmatism on its own, and with some suture removals. The differences between the values before and after the rejection episode are, therefore, likely to be primarily due to the known short-term and long-term fluctuations in the measurement of visual acuity,22 intraocular pressure23 and central corneal thickness,24 furthermore to the known intraobserver and interobserver variability in the measurement of endothelial cell number.25

Strengths and limitations

To the best of our knowledge, this is the first analysis from a prospectively planned, multicentre study focusing on the outcome after successfully reversed graft rejection.

The prospective approach had the advantage of standardised data collection with uniform risk stratification and consequently comparability across centres, and the multicentre approach ensured a sufficiently large sample size. Nevertheless, the number of missing data proved to be a major limitation of this preliminary report. On the one hand, this was due to the protocol, which specified a first standardised follow-up visit only 6 months after transplantation. As a consequence, patients with acute graft rejection within the first 6 months after transplantation could not be included in the main outcome analysis due to lack of clinical data. The same applied to patients with immunological rejection at the end of the 36-month observation period, for whom clinical values after rejection were missing. Other reasons for missing data were due to the COVID-19 pandemic (statutory cancellation of non-essential patient contacts, disease-related absences) or incomplete instrumental testing (endothelial cell density, corneal thickness). In the present data set, this consisted of 16 of 91 (18%) patients with acute graft rejection who had no follow-up before or at least 3 months after rejection and therefore could not be considered for the analysis.

A further limitation is the lack of a standardised treatment protocol after transplantation and after rejection episodes. Accordingly, no conclusion can be drawn on treatment efficacy.

Implication for clinical practice and research

The information that the visual acuity can return to baseline values after successfully overcoming rejection is relevant for the clinician and the patient: it is worth taking on the therapeutic effort and possible side effects of temporary high-dose steroid therapy.

Future studies should be optimised by standardised treatment regimens after transplantation and rejection episodes, by more frequent follow-up, especially in the first year after transplantation and an observation period longer than 36 months. External monitoring would be recommendable to ensure complete data collection in multicentre trials.

In conclusion, the present report about the outcome of corneal grafts, after reversed rejection episodes, shows that visual acuity, intraocular pressure, endothelial cell density and central corneal thickness do not change to a statistically significant and clinically relevant degree. This applies equally to PK and EK, regardless of their risk stratification.

Overview

Article metrics
Altmetric data not available for this article.
Dimensionsopen-url