Graft rejection episodes after keratoplasty in Asian eyes

doi:10.21203/rs.3.rs-2192627/v1

This study aimed to investigate the clinical characteristics and risk factors for graft rejection after keratoplasty in Asian patients. We enrolled 730 cases (566 patients) of penetrating keratoplasty (PK, N = 198), Descemet’s stripping automated endothelial keratoplasty (DSAEK, N = 277), non-Descemet’s stripping automated endothelial keratoplasty (nDSAEK, N = 138), and Descemet’s membrane endothelial keratoplasty (DMEK, N = 117). The incidence, clinical characteristics, and possible risk factors of graft rejection were analyzed. Overall, graft rejection occurred in 65 cases (56 patients, 8.9%). The incidence of rejection was highest with PK (3.45/100 person-years), followed by DSAEK (2.34), nDSAEK (1.55), and DMEK (0.24). Cox regression analysis showed that keratoplaty type, younger age, indications (such as failed keratoplasty and infection), and steroid eyedrop use were possible risk factors. Furthermore, PK had the highest hazard estimate of rejection, with significant differences between PK and the other groups (P = 0.018, DSAEK; P = 0.022, nDSAEK; and P < 0.001, DMEK). No significant difference was observed between DSAEK and nDSAEK groups (P = 0.829) in terms of graft rejecton. The DMEK group had the lowest hazard estimate of rejection, with significant differences between DMEK and the other groups (P < 0.001, PK; P = 0.006, DSAEK; and P = 0.010, nDSAEK), confirming the superiority of this procedure.

Health sciences/Medical research

Health sciences/Health care/Therapeutics/Surgery/Transplantation

Descemet’s stripping automated endothelial keratoplasty

non-Descemet’s stripping automated endothelial keratoplasty

Descemet’s membrane endothelial keratoplasty

bullous keratopathy

endothelial keratoplasty

immunological graft rejection

penetrating keratoplasty

Over the past two decades, new lamellar keratoplasties, including Descemet stripping automated endothelial keratoplasty (DSAEK) and Descemet’s membrane endothelial keratoplasty (DMEK), have been developed to avoid complications caused by conventional penetrating keratoplasty (PK).^[1–7] Additionally, we^[8] and others^{[9, 10]} eliminated Descemet stripping for bullous keratopathies in non-Fuchs type corneal endothelial dystrophy, and called the modified procedure non-Descemet Stripping Automated Endothelial Keratoplasty (nDSAEK).^[8] Moreover, both DSAEK and nDSAEK have proven quite effective for endothelial dysfunction, with rapid visual recovery and minimal induced astigmatism.^{[8, 11, 12]}

Although PK remains a viable option in the U.S., the number of DSAEK and DMEK procedures has dramatically increased in the past two decades, and the number of DSAEK and DMEK procedures was almost comparable by the end of 2021.^[13] Recently, we reported a significant reduction in PK procedures while DSAEK/nDSAEK and DMEK procedures have significantly increased in our hospital.^{[14, 15]} Although these modern lamellar keratoplasties may reduce the incidence of immunological corneal graft rejection episodes, graft rejection and subsequent graft failure still occur, even after lamellar surgeries.^[11] However, the incidence and characteristics of graft rejection after nDSAEK remain unclear.

This study aimed to investigate the clinical characteristics, risk factors, and probability of immunological graft rejection after keratoplasties, including PK and other novel lamellar keratoplasties (DSAEK/nDSAEK and DMEK), in Asian patients.

Table 1 shows the demographics and comorbidities for the total and each surgical procedure of keratoplasty (N = 730). This study included data of 730 cases (359 males, 371 females; mean age ± standard deviation [SD], 71.2 ± 10.6 years) in 566 patients that underwent PK (N = 198), DSAEK (N = 277), nDSAEK (N = 138), and DMEK (N = 117). The mean and median postoperative duration were 1559 and 1190 days, respectively (90–5758 days). Background comorbidities of all keratoplasty cases included hypertension (N = 149, 20.4%), diabetes mellitus (N = 99, 13.6%), atopic dermatitis (N = 11, 1.5%), previous herpetic keratitis (N = 42, 5.8%), pre-existing glaucoma (N = 216, 29.6%), and prior keratoplasty in the opposite eye (N = 157, 21.5%).

Table 2 summarizes the indications for the total and each surgical procedure of keratoplasty (N = 730). Failed keratoplasty was the most common indication for keratoplasty throughout the 15-year period (N = 172, 23.6%), followed by argon laser iridotomy (N = 125, 17.1%), pseudophakic bullous keratopathy (N = 82, 11.2%), glaucoma surgery (N = 78, 10.7%), corneal opacity (N = 48, 6.6%), Fuchs corneal endothelial dystrophy (N = 48, 6.6%), exfoliation syndrome (N = 27, 3.7%), keratoconus (N = 21, 2.9%), perforation (N = 17, 2.3%), infection (N = 13, 1.8%), cytomegalovirus corneal endotheliitis (N = 12, 1.6%), corneal dystrophy/degeneration (N = 11, 1.5%), iridocorneal endothelial syndrome (N = 3, 0.4%), and other causes (N = 73, 10.0%). Failed PK was the most common cause for failed keratoplasty (N = 67, 39.0%), followed by failed DSAEK (N = 67, 39.0%), failed nDSAEK (N = 17, 9.9%), failed DMEK (N = 15, 8.7%), and failed deep anterior lamellar keratoplasty (DALK) (N = 5, 2.9%). Table 2 shows the detailed indications for each keratoplasty procedure.

Table 3 summarizes the combined surgery and graft size for the total and each keratoplasty procedure (N = 730). Among all keratoplasty cases enrolled in this study, 603 cases (82.6%) were simple keratoplasty procedures, and the remaining (N = 127,17.4%) had simultaneous surgeries at the time of keratoplasty, including phacoemulsification with intraocular lens implantation and other procedures. Median graft size for PK, DSAEK, nDASEK, and DMEK were 7.75 (range: 5.5–8.75) mm, 8.0 (range: 6.0–8.5) mm, 8.0 (range: 6.0–8.5) mm, and 8.0 (range: 7.25–8.5) mm, respectively.

Table 4 summarizes the graft rejection episodes, eye drops at rejection or last visit, and graft outcomes in all cases (N = 730). Overall, graft rejection episodes occurred in 65 cases in 56 patients (8.9%). PK showed the highest rejection rate (33/198, 16.7%), followed by nDSAEK (11/138, 8.0%), DSAEK (20/277, 7.2%), and DMEK (1/117, 0.9%). The mean follow-up period for all patients (N = 730) from the time of surgery to the time of rejection or last visit (for patients without rejection) was 1468 days. At the time of rejection or last visit, the following 627 cases (85.9%) used steroid eye drops: PK (N = 163, 82.3%), DSAEK (N = 246, 88.8%), nDSAEK (N = 103, 74.6%), and DMEK (N = 115, 98.3%). Among steroid eye drops administered, betamethasone was used in 164 cases (26.2%), and fluorometholone was used in the remaining 463 cases (73.8%). Regarding the graft outcomes in all cases (N = 730), 574 cases (78.6%) remained clear, while the remaining cases (N = 156, 21.4%) failed. Subgroup analysis of the surgical procedure of keratoplasty and clear graft without rejection were observed in PK (N = 111, 56.1%), DSAEK (N = 204, 73.6%), nDSAEK (N = 101, 73.2%), and DMEK (N = 108, 92.3%).

Table 5 summarizes findings on rejection episodes and graft outcomes in limited cases with a rejection episode (N = 65). The mean onset of rejection after overall keratoplasty was 641 ± 670 (days ± SD): PK (717 ± 735 days), DSAEK (507 ± 460 days), nDSAEK (661 ± 829 days), and DMEK (629 days). Figure 1 shows representative slit-lamp photographs of the graft rejection for each keratoplasty procedure. As shown in Table 5, 17 of 33 PK rejection cases (51.5%), 8 of 20 DSAEK rejection cases (40.0%), 7 of 11 nDSAEK rejection cases (63.6%), and 1 of 1 DMEK rejection case (100%) were symptomatic. Signs of immunological rejection at the initial diagnosis in PK, DSAEK, nDSAEK, and DMEK included conjunctival hyperemia (27.3%, 30.0%, 18.2%, and 0%, respectively), diffuse corneal edema (45.5%, 45.0%, 45.5%, and 100%, respectively), and keratic precipitates (84.8%, 95.0%, 90.9%, and 100%, respectively). Although this study demonstrated rejection lines in 4 cases of PK, no endothelial rejection lines were observed in all types of endothelial keratoplasty. Most rejected grafts cleared after the episode (50, 76.9%); however, 10 eyes with rejection (30.3%) that underwent PK, 3 eyes with rejection (15.0%) that underwent DSAEK, and 2 eyes with rejection (18.2%) that underwent nDSAEK progressed to graft failure.

Table 6 summarizes the incidence rates of rejection episodes according to the time interval after each type of keratoplasty (N = 730). The overall incidence rates of rejection episodes [per 100 person-years] were the highest in PK (3.45), followed by DSAEK (2.34), nDSAEK (1.55), and DMEK (0.24) in descending order. In the DMEK group, only one rejection episode was observed, and the overall estimate of the rejection incidence rate was extremely low compared with other types of keratoplasty. In the other three groups, PK, DSAEK, and nDSAEK, the highest incidence rate was observed between 0 and 1 year after each keratoplasty.

Figure 2 shows the Kaplan–Meier curve of rejection episodes by surgical procedure group (N = 730). The PK group had the highest hazard estimate of rejection, with significant differences between this group and the other three groups (P = 0.018, DSAEK; P = 0.022, nDSAEK; and P < 0.001, DMEK). Additionally, no significant difference was observed between DSAEK and nDSAEK (HR = 0.92, 95% CI = [0.44, 1.95], P = 0.829). The DMEK group had the lowest hazard estimate of rejection, and significant differences were observed between this group and the other three groups (P < 0.001, PK; P = 0.006, DSAEK; and P = 0.010, nDSAEK).

Table 7 shows the results of the univariate and multivariate Cox regression analyses of rejection episodes (N = 730). In the univariate model, surgical procedure, age, indications (failed keratoplasty and infection), and eye drop at rejection had significant effects on rejection episodes (P < 0.001, 0.005, 0.014, 0.021, and < 0.001, respectively). In the multivariate model with baseline characteristics, the surgical procedure, age, and indications (failed keratoplasty and infection) remained after backward variable selection. Moreover, younger age, indication for failed keratoplasty, and infection were identified as risk factors at baseline (P = 0.023, 0.019, and 0.043, respectively). In the multivariate model with eye drops, significant effects of eye drops at rejection or last visit (P = 0.019) were observed after adjustment for baseline risk factors. Compared to eye drop non-users at rejection or last visit, a higher risk of rejection was observed in eye drop users (fluorometholone HR:1.37, 95% CI: [0.63, 3.00] and betamethasone HR:2.68, 95% CI: [1.19, 6.04] ).

The results of subgroup analyses on rejection episodes by demographics, indication for keratoplasty, and simple/combined keratoplasty and graft size that correspond to Tables 1, 2, and 3 are summarized in Supplementary Tables 1 to 3 (Tables S1, S2, S3).

This study investigated the clinical characteristics, risk factors, and probability of immunological graft rejection after keratoplasty, including PK, DSAEK, nDSAEK, and DMEK in Asian patients. Therefore, this study showed graft rejection episodes occurred in 65 of the 56 patients (65/730, 8.9%) who underwent keratoplasty. In the assessment of each keratoplasty technique, PK had the highest rejection rate (33/198, 16.7%) than other types of surgeries, followed by nDSAEK (11/138, 8.0%), DSAEK (20/277, 7.2%), and DMEK (1/117, 0.9%) (Table 4). For improved accuracy in statistical analysis, the incidence rates of rejection episodes were investigated using the person-years method. Consequently, similar outcomes were obtained; PK was the highest (3.45/100 person-years), followed by DSAEK (2.34), nDSAEK (1.55), and DMEK (0.24) in descending order (Table 6). Furthermore, Cox regression analysis of the hazard ratio between surgical groups showed that the PK group had the highest hazard estimate of rejection, and significant differences were observed between this group and the other three groups (P = 0.018, DSAEK; P = 0.022, nDSAEK; and P < 0.001, DMEK). Most notably, no significant difference was observed between DSAEK and nDSAEK (HR = 0.92, 95% CI = [0.44, 1.95], P = 0.829). The DMEK group had the lowest hazard estimate of rejection, and significant differences were observed between this group and the other three groups (P < 0.001, PK; P = 0.006, DSAEK; and P = 0.010, nDSAEK). These results were consistent with those of previous studies,^[16-21] thus confirming the excellent advantages of a lower rejection rate of endothelial keratoplasties compared with conventional PK; DMEK,in particular, showed the lowest incidence rate of graft rejection, clearly indicating the superiority of this procedure.^[16] To the best of our knowledge, this study is the largest case series of recent keratoplasties in Japan.

The reasons for lower graft rejection after endothelial keratoplasty are as follows:^[22] (a) The transplanted tissue is inserted into the anterior chamber and has no exposure to the surface, where antigen-presenting cells and antibodies are present. Hence, anterior chamber-associated immune deviation may be effective. (b) A significant reduction in the number of sutures connecting the host and donor tissues may lead to fewer suture-related rejection episodes. (c) The absence of direct contact between the host stromal vessels and transplanted tissue disrupts the immune effector and effector arcs. (d) Reduced immunogenicity of the donated tissue due to the absence of donor epithelium and the majority of the stroma.

Clinical outcomes of nDSAEK have been reported to be similar to those of DSAEK.^{[8,11,12,23-29]} Several reports support the advantages of nDSAEK, particularly for failed PK, as stripping Descemet’s membrane of failed PK may cause PK wound separation or weakness.^[23-25] The histology and in vivo observation of the residual host endothelial cells after nDSAEK have already been thoroughly studied, including decreased density, loss of pump function, apoptosis, and altered morphology in a rabbit model by in vivo confocal microscopy.^[30] Additionally, in vivo laser confocal microscopy in human studies after nDASEK can detect subclinical corneal abnormalities, including subepithelial haze, host-recipient interface haze, host stromal needle-shaped materials, and host-recipient interface particles with characteristic giant particles.^[31] Using immunohistochemistry and transmission electron microscopy, 2 weeks after nDSAEK in a rabbit model, host endothelial cells appeared to be morphologically altered, occasionally detached from the adjacent Descemet membrane, extending into the graft stroma or engulfing strands of the grafted stroma at the interface.^[32] However, the rate of graft rejection after nDSAEK has not yet been reported. Herein, we reported for the first time that the rate of graft rejection following nDSAEK was almost identical to that after DSAEK by a hazard ratio similar to that of the multivariate model without eye drops (DSAEK: 0.57, nDSAEK: 0.56) (Table 7). This result was also confirmed by the group comparison shown in Figure 1 (DSAEK vs. nDSAEK, P = 0.829), indicating that the presence of host Descemet’s membrane showed no difference between nDSAEK and DSAEK in terms of rejection.

In this study, signs of immunological rejection by slit-lamp biomicroscopic observation at diagnosis in PK, DSAEK, nDSAEK, and DMEK included conjunctival hyperemia (27.3%, 30.0%, 18.2%, and 0%, respectively), diffuse corneal edema (45.5%, 45.0%, 45.5%, and 100%, respectively), and keratic precipitates (84.8%, 95.0%, 90.9%, and 100%, respectively) (Table 5). A rejection line was observed in four cases of PK; however, no endothelial rejection lines were observed in other types of keratoplasty. Contrarily, the symptomatic rejection rates in PK, DSAEK, nDSAEK, and DMEK were 51.5%, 40.0%, 63.6%, and 100%, respectively. Collectively, careful observation using slit lamp biomicrography to detect corneal edema and keratic precipitates is crucial in detecting the early phase of graft rejection. Routine examination of corneal thickness using pachymetry and anterior segment optical coherence tomography may also help to detect subclinical graft rejection. Considering half of the rejection cases were asymptomatic, ophthalmologists should look out for signs of rejection during routine postoperative examinations.

In this study, we investigated the risk factors of graft rejection (Table 7). In the univariate model, surgical procedure, age, indications (failed keratoplasty and infection), and eye drops at the observational endpoint (rejection or last visit) had significant effects on rejection episodes, of which the most influential factor was the surgery type. The multivariate model with baseline characteristics without considering steroid eye drops usage, surgical procedure, younger age, and indications (failed keratoplasty and infection) remained after backward variable selection. In the multivariate model with eye drops (i.e., considering steroid eye drops usage), significant effects of eye drop at rejection or last visit (P = 0.019) were observed; steroid eye drop users had a significant risk of graft rejection compared to non-users of eye drops (i.e., fluorometholone HR: 1.37, 95% CI: [0.63, 3.00] and betamethasone HR: 2.68, 95% CI: [1.19, 6.04]). According to the postoperative treatment regimen in the methods section, topical 0.1% betamethasone was changed to a soft steroid, such as 0.1% fluorometholone, thrice a day after 6 months. By reviewing the chart of the cases with rejection episodes, we observed use of topical 0.1% betamethasone in the following 25 cases (38.5%): PK (N = 18, 54.5%), DSAEK (N = 5, 25.0%), nDASEK (N = 2, 18.2%), and DMEK (N = 0) (Table 5). Of these, 9 cases (36%) had a history of failed keratoplasty. Furthermore, another 4 cases (16%) had a history of eye surgery other than cataract surgery (e.g., vitrectomy). We hypothesized that these complicated cases had a higher risk of rejection on account of having been treated with stronger steroid eye drops such as 0.1% betamethasone. Other factors, including sex, hypertension, diabetes mellitus, herpetic keratitis, glaucoma with or without surgery, prior keratoplasty in the opposite eye, indications for keratoplasty except for failed keratoplasty and infection, combined surgery, and graft size were not identified as risk factors for graft rejection.

One limitation of this study was the difference in the background diseases of keratoplasty between Japan and other countries. In Western countries, Fuchs corneal endothelial dystrophy accounts for most of the cases of endothelial dysfunction; however, in our study, failed keratoplasty (23.6%), argon laser iridotomy (17.1%), pseudophakic bullous keratopathy (11.2%), and glaucoma surgery (10.7%) were the leading causes, and Fuchs corneal endothelial dystrophy (6.6%) was relatively rare. Nonetheless, the information in this study is the largest keratoplasty case series in Japan and may contribute to the understanding of graft rejection in keratoplasty, which remains a severe challenge for ophthalmologists.

In conclusion, our study revealed that the type of keratoplasty was the most significant factor for graft rejection in our cohort. PK showed the highest rejection rate, followed by nDSAEK and DSAEK. Additionally, DMEK showed a relatively low incidence rate of graft rejection, confirming the superiority of this procedure over other types of keratoplasty. The presence of the host Descemet’s membrane showed no difference between the nDSAEK and DSAEK groups in terms of rejection episodes. Considering that patients who are asymptomatic exist in about half of the rejection cases, ophthalmologists should be aware of the diagnosis of rejection episodes during routine postoperative examinations.

Study design and patient approval

This retrospective chart review was approved by the Ethical Committee of Kanazawa University Graduate School of Medical Science (approval number: 3483-1), and the work was compliant with the Health Insurance Portability and Accountability Act of 1996 and consistent with Good Clinical Practices. It adhered to the tenets of the Declaration of Helsinki, and written informed consent was obtained from all the patients.

Anonymized data from consecutive cases of all types of corneal transplantation, including PK, anterior lamellar keratoplasty (ALK), DALK, DSAEK, nDSAEK, and DMEK, in two participating centers in Japan (Kanazawa University Hospital and Yokohama Minami Kyosai Hospital) between January 2006 and December 2020 were collected. However, cases of ALK (N = 6) and DALK (N = 33) were excluded from this study; statistical analysis was impossible due to the small sample size. Cases with patients younger than 40 years old were excluded from this study because these cases were relatively rare in Japan and sometimes quite complicated with multiple surgical histories, which may lead to misinterpretation of the whole data. Cases without chart information regarding the type of eye drop at the time of rejection or the final follow-up visit were excluded. Additionally, cases with less than 3 months (90 days) of follow-up data were excluded, leaving 730 cases for the analysis of 566 patients. The medical records of these patients were reviewed. Notably, the information including recipient age, sex, diagnosis, types of keratoplasties, surgical histories, duration of follow-up, simultaneous surgeries at the time of keratoplasty, systemic comorbidities (hypertension, diabetes, and atopic dermatitis), ophthalmic histories (previous herpetic keratitis, pre-existing glaucoma, and prior keratoplasty in opposite eyes), clinical symptoms, final outcome regarding failure or success of the graft, and types of steroidal medications at the time of rejection or final follow-up visit were analyzed. Moreover, slit-lamp presentations, including conjunctival hyperemia, corneal edema, keratic precipitates, vascularization of the cornea, peripheral anterior synechia, and rejection line, were recorded for patients with graft rejection. Graft failure was defined as an irreversible loss of central corneal clarity with stromal thickening from any cause with or without an endothelial rejection line, inflammation including keratic precipitates, cells in the stroma, and an increase in aqueous cells from a previous visit.

Surgical techniques

Three trained surgeons performed the keratoplasty surgeries at Kanazawa University Hospital (AK and HY) and Yokohama Minami Kyosai Hospital (TH). PK was performed conventionally with a 10–0 nylon suture using either a running suture or an interrupted suture technique. In most cases, a corneal trephine with a diameter of 7.5 mm with a 7.75 mm donor punch was used for PK. DSAEK was performed as previously described using a double glide (Busin glide (Moria, Antony, France) and intraocular lens sheet glide) donor insertion technique^[33] or NS Endo-Inserter (HOYA Co., Ltd, Tokyo Japan).^[34] Domestic eye bank donors dissected with a microkeratome (ALTK Cbm, Moria Japan KK, Tokyo, Japan) or precut DSAEK donors internationally shipped from a US eye bank (CorneaGen Inc., Seattle, WA, USA) with a diameter of 8.0 mm were used in most cases. For nDSAEK, neither Descemet’s membrane scoring in a circular pattern nor Descemet’s membrane stripping was performed. DMEK was performed as previously described using a pre-stripped pre-punched pre-loaded donor from a US eye bank (CorneaGen Inc., Seattle, WA, USA) with a diameter of 8.0 mm.^[35-39]

Postoperative treatment regimen　

Postoperatively, topical 0.5% levofloxacin and topical 0.1% betamethasone were applied five times daily for 1 week, then tapered to four times daily for 6 months, provided they did not indicate steroid-induced glaucoma. After 6 months, topical 0.1% betamethasone was changed to soft steroids, including 0.1% fluorometholone, thrice daily. Moreover, following 1 year, the eyes were usually maintained on once-daily steroid dosing indefinitely as long as no signs or symptoms of graft rejection occurred. In eyes with steroid-induced intraocular pressure (IOP) control problems, topical 0.1% betamethasone was usually changed to a soft steroid, such as 0.1% fluorometholone, and dosing was adjusted as needed to adequately lower IOP. Glaucoma agents were added if necessary. For cases of endothelial keratoplasties, including DSAEK, nDSAEK, and DMEK, 0.1% bromfenac eyedrop 2 times daily was used for 1 month to prevent postoperative cystoid macular edema.

At the time of a rejection episode, topical 0.1% betamethasone was prescribed 8 times daily, approximately hourly while awake and oral prednisolone was added in cases where graft rejection seemed severe. The dosage of topical 0.1% betamethasone was maintained for 1 week and gradually tapered to 3–4 times a day for several months.

Statistical analysis

Demographics and other baseline characteristics were summarized with appropriate descriptive statistics such as mean (SD) by the surgical procedure group, PK, DSAEK, nDSAEK, and DMEK groups. Eye drops at rejection or the last visit and graft outcomes were also summarized. In this analysis, the date rejection was first diagnosed was recorded as the date of the “event” or eyes that did not experience a rejection episode, and the date of the most recent follow-up exam with no evidence of rejection was used to determine the length of follow-up. The findings and graft outcomes were summarized for cases with a rejection episode. Additionally, the incidence rate of rejection was calculated as the time interval after keratoplasty using the person-year method. The time to rejection from keratoplasty was graphically described by the surgical procedure group using the Kaplan–Meier curve. Hazard ratios between the groups were estimated using Cox regression, and the curves were compared using the log-rank test. Univariate and multivariate Cox proportional hazards regressions were applied to detect risk factors for rejection episodes. Moreover, the multivariate model with baseline characteristics included surgical procedure, age, sex, hypertension, diabetes mellitus, herpetic keratitis, glaucoma with or without surgery, prior keratoplasty in the opposite eye, indications for keratoplasty, combined surgery, and graft size as exploratory variables, which were selected by backward elimination. Eye drops at rejection or the last visit were included in the multivariate model with baseline characteristics.

Statistical significance was set at P < 0.05. Statistical analyses were performed using R, version 4.2.1.^[40]

Data availability

The authors declare that all data generated or analyzed during this study are included in this published article and its supplementary information files.

Acknowledgements: We thank Eisei Oda, PhD, of StatLink Consulting, for providing statistical consultation on the data analyses.

Author contributions: H.W.: data collection and writing; T.H.: analysis and investigation; A.K.: review and editing, writing, and project administration; T.N.: data collection and curation; N.M. and H.Y.: investigation; S.Y.: review and supervision and validation; and K.S.: supervision and validation.

All authors critically checked the manuscript and approved the submission.

Competing interest: The authors declare no competing interests.

Funding support: This work did not receive any specific grants from funding agencies in the public, commercial, or not-for-profit sectors.

Tan, D. T. H., Dart, J. K. G., Holland, E. J. & Kinoshita, S. Corneal transplantation. Lancet 379, 1749–1761 (2012).
Anshu, A., Price, M. O., Tan, D. T. H. & Price, F. W. Endothelial keratoplasty: a revolution in evolution. Surv. Ophthalmol. 57, 236–252 (2012).
Melles, G. R. J. et al. A surgical technique for posterior lamellar keratoplasty. Cornea 17, 618–626 (1998).
Price, F. W. & Price, M. O. Descemet’s stripping with endothelial keratoplasty in 200 eyes: early challenges and techniques to enhance donor adherence. J. Cataract Refract. Surg. 32, 411–418 (2006).
Gorovoy, M. S. Descemet-stripping automated endothelial keratoplasty. Cornea 25, 886–889 (2006).
Melles, G. R. J., Lander, F. & Rietveld, F. J. R. Transplantation of Descemet’s membrane carrying viable endothelium through a small scleral incision. Cornea 21, 415–418 (2002).
Melles, G. R. J., Ong, T. S., Ververs, B. & Van Der Wees, J. Descemet membrane endothelial keratoplasty (DMEK). Cornea 25, 987–990 (2006).
Kobayashi, A., Yokogawa, H. & Sugiyama, K. Non-Descemet stripping automated endothelial keratoplasty for endothelial dysfunction secondary to argon laser iridotomy. Am. J. Ophthalmol. 146, 543–549 (2008).
Price, M. O., Price, F. W. & Trespalacios, R. Endothelial keratoplasty technique for aniridic aphakic eyes. J. Cataract Refract. Surg. 33, 376–379 (2007).
Price, F. W. & Price, M. O. Endothelial keratoplasty to restore clarity to a failed penetrating graft. Cornea 25, 895–899 (2006).
Yamazaki, R. et al. Descemet stripping and automated endothelial keratoplasty (DSAEK) versus non-Descemet stripping and automated endothelial keratoplasty (nDSAEK) for bullous keratopathy. Jpn. J. Ophthalmol. 64, 585–590 (2020).
Omoto, T. et al. Comparison of 5-year clinical results of Descemet and non-Descemet stripping automated endothelial keratoplasty. Cornea 39, 573–577 (2020).
Eye Bank Association of America, (2021). Eye Banking Statistical Report http://www.restoresight.org/.
Nishino, T., Kobayashi, A., Yokogawa, H., Mori, N. & Sugiyama, K. Changing indications and surgical techniques for keratoplasty during a 16-year period (2003–2018) at a tertiary referral hospital in Japan. Clin. Ophthalmol. 13, 1499–1509 (2019).
Nishino, T. et al. A 10-year review of underlying diseases for endothelial keratoplasty (DSAEK/DMEK) in a tertiary referral hospital in Japan. Clin. Ophthalmol. 12, 1359–1365 (2018).
Hos, D. et al. Immune reactions after modern lamellar (DALK, DSAEK, DMEK) versus conventional penetrating corneal transplantation. Prog. Retin. Eye Res. 73, 100768 (2019).
Debourdeau, E. et al. Risk factors of rejection after penetrating keratoplasty: a retrospective monocentric study. Graefes Arch. Clin. Exp. Ophthalmol. (2022).
Jordan, C. S., Price, M. O., Trespalacios, R. & Price, F. W. Graft rejection episodes after Descemet stripping with endothelial keratoplasty: part one: clinical signs and symptoms. Br. J. Ophthalmol. 93, 387–390 (2009).
Price, M. O., Jordan, C. S., Moore, G. & Price, F. W. Graft rejection episodes after Descemet stripping with endothelial keratoplasty: part two: the statistical analysis of probability and risk factors. Br. J. Ophthalmol. 93, 391–395 (2009).
Allan, B. D. S. et al. Corneal transplant rejection rate and severity after endothelial keratoplasty. Cornea 26, 1039–1042 (2007).
Anshu, A., Price, M. O. & Price, F. W. Risk of corneal transplant rejection significantly reduced with Descemet’s membrane endothelial keratoplasty. Ophthalmology 119, 536–540 (2012).
Prakash, G., Jhanji, V. & Titiyal, J. S. Will Descemet’s stripping with automated endothelial keratoplasty (DSAEK) lower the rates of allograft rejection in corneal transplants for endothelial failure? Med. Hypo. 69, 1117–1119 (2007).
Chaurasia, S., Ramappa, M., Murthy, S. I., Garg, P. & Sangwan, V. S. Endothelial keratoplasty without stripping the Descemet’s membrane. Br. J. Ophthalmol. 95, 1473–1474 (2011).
Nottage. JM Nirankari, V. S. Endothelial keratoplasty without Descemet’s stripping in eyes with previous penetrating corneal transplants. Br. J. Ophthalmol. 96, 24–27 (2012).
Heitor De Paula, F., Kamyar, R., Shtein, R. M., Sugar, A. & Mian, S. I. Endothelial keratoplasty without Descemet stripping after failed penetrating keratoplasty. Cornea 31, 645–648 (2012).
Masaki, T., Kobayashi, A., Yokogawa, H., Saito, Y. & Sugiyama, K. Clinical evaluation of non-Descemet stripping automated endothelial keratoplasty (nDSAEK). Jpn. J. Ophthalmol. 56, 203–207 (2012).
Minezaki, T., Hattori, T., Nakagawa, H., Kumakura, S. & Goto, H. Non-Descemet’s stripping automated endothelial keratoplasty for bullous keratopathy secondary to iridoschisis. Clin. Ophthalmol. 7, 1353–1355 (2013).
Yokogawa, H. et al. Descemet’s stripping and non-Descemet’s stripping automated endothelial keratoplasty for microcornea using 6.0 mm donor grafts. Clin. Ophthalmol. 7, 1951–1956 (2013).
Ren, Y. et al. Viscoelastic-assisted non-Descemet stripping automated endothelial keratoplasty in vitrectomized and iris-lens diaphragm injured eyes. Eye Contact Lens 41, 398–402 (2015).
Sun, J. P., Hu, F. R., Chen, Y. M., Chu, H. S. & Chen, W. L. Change of recipient corneal endothelial cells after non-Descemet’s stripping automated endothelial keratoplasty in a rabbit model. Invest. Ophthalmol. Vis. Sci. 55, 8467–8474 (2014).
Kobayashi, A., Yokogawa, H. & Sugiyama, K. In vivo laser confocal microscopy after non-Descemet’s stripping automated endothelial keratoplasty. Ophthalmology 116, 1306–1313 (2009).
Hatanaka, H. et al. A study of host corneal endothelial cells after non-Descemet stripping automated endothelial keratoplasty. Cornea 32, 76–80 (2013).
Kobayashi, A., Yokogawa, H. & Sugiyama, K. Descemet stripping with automated endothelial keratoplasty for bullous keratopathies secondary to argon laser iridotomy--preliminary results and usefulness of double-glide donor insertion technique. Cornea 27 Suppl 1, S62–S69 (2008).
Kobayashi, A., Yokogawa, H., Mori, N., Nishino, T. & Sugiyama, K. Clinical evaluation of the NS Endo-inserter, a novel donor inserter for Descemet’s stripping automated endothelial keratoplasty. Case Rep. Ophthalmol. 10, 357–364 (2019).
Monnereau, C. et al. Multicenter study of Descemet membrane endothelial keratoplasty: first case series of 18 surgeons. JAMA Ophthalmol. 132, 1192–1198 (2014).
Kobayashi, A. et al. Evaluation of internationally shipped prestripped donor tissue for Descemet membrane endothelial keratoplasty by vital dye staining. Cornea 34, 225–227 (2015).
Kobayashi, A., Yokogawa, H., Yamazaki, N., Masaki, T. & Sugiyama, K. In vivo laser confocal microscopy after Descemet’s membrane endothelial keratoplasty. Ophthalmology 120, 923–930 (2013).
Kobayashi, A., Yokogawa, H., Yamazaki, N., Masaki, T. & Sugiyama, K. The use of endoillumination probe-assisted Descemet membrane endothelial keratoplasty for bullous keratopathy secondary to argon laser iridotomy. Clin. Ophthalmol. 9, 91–93 (2015).
Kobayashi, A., Yokogawa, H., Mori, N., Nishino, T. & Sugiyama, K. Graft edge reflection of a tightly scrolled roll using endoillumination as a simple method for determining graft orientation in Descemet membrane endothelial keratoplasty. Cornea 40, 254–257 (2021).
R Core Team. R: A Language and Environment for Statistical Computing. https://www.R-project.org/ (R Foundation for Statistical Computing, Vienna, Austria, 2022).

Table 1. Demographics and comorbidities by surgical procedure of keratoplasty

Characteristic	Total N = 730	Surgical procedure of keratoplasty
Characteristic	Total N = 730	[A] PK N = 198	[B] DSAEK N = 277	[C] nDSAEK N = 138	[D] DMEK N = 117
Number of patients, n	566	164	227	123	108
Age [years]
Mean (SD)	71.2 (10.6)	67.2 (12.3)	72.5 (9.8)	73.1 (10.2)	72.6 (8.1)
Median [Range]	73.0 [41.0, 97.0]	70.0 [41.0, 90.0]	74.0 [41.0, 97.0]	74.0 [45.0, 90.0]	74.0 [48.0, 89.0]
Sex, n (%)
[1] Male	359 (49.2%)	119 (60.1%)	145 (52.3%)	56 (40.6%)	39 (33.3%)
[2] Female	371 (50.8%)	79 (39.9%)	132 (47.7%)	82 (59.4%)	78 (66.7%)
Hypertension, n (%)	149 (20.4%)	32 (16.2%)	69 (24.9%)	36 (26.1%)	12 (10.3%)
Diabetes mellitus, n (%)	99 (13.6%)	21 (10.6%)	52 (18.8%)	18 (13.0%)	8 (6.8%)
Atopic dermatitis, n (%)	11 (1.5%)	5 (2.5%)	6 (2.2%)	0 (0.0%)	0 (0.0%)
Herpetic keratitis, n (%)	42 (5.8%)	27 (13.6%)	9 (3.2%)	5 (3.6%)	1 (0.9%)
Glaucoma and surgery, n (%)
[1] No glaucoma	514 (70.4%)	150 (75.8%)	158 (57.0%)	98 (71.0%)	108 (92.3%)
[2] Glaucoma without surgery	64 (8.8%)	14 (7.1%)	27 (9.7%)	15 (10.9%)	8 (6.8%)
[3] Glaucoma with surgery	152 (20.8%)	34 (17.2%)	92 (33.2%)	25 (18.1%)	1 (0.9%)
Prior keratoplasty in opposite eye, n (%)
[1] No	572 (78.4%)	158 (79.8%)	222 (80.1%)	113 (81.9%)	79 (67.5%)
[2] Yes	157 (21.5%)	39 (19.7%)	55 (19.9%)	25 (18.1%)	38 (32.5%)
[3] Unknown	1 (0.1%)	1 (0.5%)	0 (0.0%)	0 (0.0%)	0 (0.0%)
Follow-up period up to last visit [days]
Mean (SD)	1559 (1197)	1949 (1446)	1173 (852)	1992 (1,399)	1307 (714)
Median [Range]	1190 [90, 5758]	1490 [97, 5758]	948 [90, 5632]	1854 [95, 5198]	1123 [165, 3856]

DMEK, Descemet’s membrane endothelial keratoplasty; (n)DSAEK, (Non-)Descemet’s stripping automated endothelial keratoplasty; PK, penetrating keratoplasty.

Table 2. Indication of keratoplasty

Characteristic	Total N = 730	Surgical procedure of keratoplasty
Characteristic	Total N = 730	[A] PK N = 198	[B] DSAEK N = 277	[C] nDSAEK N = 138	[D] DMEK N = 117
Indication, n (%)
[01] Failed keratoplasty	172 (23.6%)	44 (22.2%)	95 (34.3%)	26 (18.8%)	7 (6.0%)
[02] ALI	125 (17.1%)	4 (2.0%)	33 (11.9%)	47 (34.1%)	41 (35.0%)
[03] PBK	82 (11.2%)	6 (3.0%)	34 (12.3%)	26 (18.8%)	16 (13.7%)
[04] Glaucoma surgery	78 (10.7%)	14 (7.1%)	48 (17.3%)	16 (11.6%)	0 (0.0%)
[05] Corneal opacity	48 (6.6%)	40 (20.2%)	7 (2.5%)	0 (0.0%)	1 (0.9%)
[06] FED	48 (6.6%)	0 (0.0%)	17 (6.1%)	3 (2.2%)	28 (23.9%)
[07] XFS	27 (3.7%)	1 (0.5%)	11 (4.0%)	6 (4.3%)	9 (7.7%)
[08] Keratoconus	21 (2.9%)	21 (10.6%)	0 (0.0%)	0 (0.0%)	0 (0.0%)
[09] Perforation	17 (2.3%)	17 (8.6%)	0 (0.0%)	0 (0.0%)	0 (0.0%)
[10] Infection	13 (1.8%)	13 (6.6%)	0 (0.0%)	0 (0.0%)	0 (0.0%)
[11] CMV corneal endotheliitis	12 (1.6%)	0 (0.0%)	7 (2.5%)	1 (0.7%)	4 (3.4%)
[12] Corneal dystrophy/degeneration	11 (1.5%)	7 (3.5%)	1 (0.4%)	0 (0.0%)	3 (2.6%)
[13] ICE	3 (0.4%)	1 (0.5%)	2 (0.7%)	0 (0.0%)	0 (0.0%)
[14] Others	73 (10.0%)	30 (15.2%)	22 (7.9%)	13 (9.4%)	8 (6.8%)
Failed keratoplasty breakdown, n
[1] Failed PK	67	36	14	15	2
[2] Failed DSAEK	67	5	58	4	0
[3] Failed nDSAEK	17	0	9	6	2
[4] Failed DMEK	15	0	12	0	3
[5] Failed DALK	5	3	2	0	0
[6] Unknown	1	0	0	1	0

Abbreviations: ALI, argon laser iridotomy; CMV, cytomegalovirus; DALK, Deep anterior lamellar keratoplasty; DMEK, Descemet’s membrane endothelial keratoplasty; (n)DSAEK, (non-)Descemet’s stripping automated endothelial keratoplasty; FED, Fuchs’ endothelial dystrophy; ICE, iridocorneal endothelial syndrome; PBK, pseudophakic bullous keratopathy; PK, penetrating keratoplasty; XFS, exfoliation syndrome.

Table 3. Combined surgery and graft size of keratoplasty

Characteristic	Total N = 730	Surgical procedure of keratoplasty
Characteristic	Total N = 730	[A] PK N = 198	[B] DSAEK N = 277	[C] nDSAEK N = 138	[D] DMEK N = 117
Simple vs. Combined keratoplasty, n (%)
[1] Simple keratoplasty	603 (82.6%)	160 (80.8%)	230 (83.0%)	113 (81.9%)	100 (85.5%)
[2] Combined keratoplasty*	127 (17.4%)	38 (19.2%)	47 (17.0%)	25 (18.1%)	17 (14.5%)
Combined surgery breakdown, n
[1] Cataract surgery^♰	102	27	35	23	17
[2] IOL fixaction	20	8	10	2	0
[3] Anterior vitrectomy	2	2	0	0	0
[4] Pupiloplasty	2	0	2	0	0
[5] KLAL	1	1	0	0	0
Graft size [mm]
Mean (SD)	7.77 (0.44)	7.63 (0.42)	7.79 (0.50)	7.80 (0.43)	7.95 (0.25)
Median [Range]	8.00 [5.50, 8.75]	7.75 [5.50, 8.75]	8.00 [6.00, 8.50]	8.00 [6.00, 8.50]	8.00 [7.25, 8.50]
Graft size group, n (%)
[1] 4.0-<7.6 mm	195 (26.7%)	66 (33.3%)	79 (28.5%)	36 (26.1%)	14 (12.0%)
[2] 7.6-<7.8 mm	128 (17.5%)	105 (53.0%)	4 (1.4%)	5 (3.6%)	14 (12.0%)
[3] 7.8-<8.1 mm	316 (43.3%)	18 (9.1%)	155 (56.0%)	80 (58.0%)	63 (53.8%)
[4] 8.1-<9.0 mm	57 (7.8%)	5 (2.5%)	27 (9.7%)	6 (4.3%)	19 (16.2%)
[5] Unknown	34 (4.7%)	4 (2.0%)	12 (4.3%)	11 (8.0%)	7 (6.0%)

DMEK, Descemet’s membrane endothelial keratoplasty; (n)DSAEK, (Non-)Descemet’s stripping automated endothelial keratoplasty; KLAL, Keratolimbal allograft; PK, penetrating keratoplasty.
* For cases with multiple concomitant techniques, the most representative technique was selected.
♰ ECCE (extracapsular cataract extraction), ICCE (intracapsular cataract extraction) and/or IOL (intraocular lens)

Table 4. Eye drop at rejection (or last visit) and graft outcome after keratoplasty in all cases (N = 730)

Characteristic	Total N = 730	Surgical procedure of keratoplasty
Characteristic	Total N = 730	[A] PK N = 198	[B] DSAEK N = 277	[C] nDSAEK N = 138	[D] DMEK N = 117
Cases with rejection episode, n (%)	65 (8.9%)	33 (16.7%)	20 (7.2%)	11 (8.0%)	1 (0.9%)
Patients with rejection episode, n Follow-up period for all cases (rejection or last visit) [days]	56	29	20	10	1
Mean (SD)	1468 (1184)	1764 (1460)	1126 (860)	1875 (1386)	1295 (714)
Median [Range]	1138 [5, 5758]	1301 [26, 5758]	888 [5, 5632]	1728 [95, 5198]	1115 [165, 3856]
Eye drop usage at rejection (or last visit), n (%)
[1] No	103 (14.1%)	35 (17.7%)	31 (11.2%)	35 (25.4%)	2 (1.7%)
[2] Yes	627 (85.9%)	163 (82.3%)	246 (88.8%)	103 (74.6%)	115 (98.3%)
Eye drop breakdown, n
[1] Fluorometholone	463	103	168	82	110
[2] Betamethasone	164	60	78	21	5
Graft outcome in all cases, n (%)
[1] Clear	524 (71.8%)	111 (56.1%)	204 (73.6%)	101 (73.2%)	108 (92.3%)
[2] Recovered	50 (6.8%)	23 (11.6%)	17 (6.1%)	9 (6.5%)	1 (0.9%)
[3] Failed	156 (21.4%)	64 (32.3%)	56 (20.2%)	28 (20.3%)	8 (6.8%)

DMEK, Descemet’s membrane endothelial keratoplasty; (n)DSAEK, (Non-)Descemet’s stripping automated endothelial keratoplasty; PK, penetrating keratoplasty.

Table 5. Findings at rejection episode and graft outcome after the episode [Cases with rejection episode] (N = 65)

Characteristic	Cases with rejection N = 65	Surgical procedure of keratoplasty
Characteristic	Cases with rejection N = 65	[A] PK N = 33	[B] DSAEK N = 20	[C] nDSAEK N = 11	[D] DMEK N = 1
Time of onset from keratoplasty [days]
Mean (SD)	641 (670)	717 (735)	507 (460)	661 (829)	629 (NA)
Median [Range]	410 [5, 3149]	438 [26, 3149]	284 [5, 1408]	362 [132, 2994]	629 [629, 629]
Eye drop, n (%)
[1] No	8 (12.3%)	1 (3.0%)	4 (20.0%)	3 (27.3%)	0 (0.0%)
[2] Fluorometholone	32 (49.2%)	14 (42.4%)	11 (55.0%)	6 (54.5%)	1 (100.0%)
[3] Betamethasone	25 (38.5%)	18 (54.5%)	5 (25.0%)	2 (18.2%)	0 (0.0%)
Suture related event, n (%)	12 (18.5%)	8 (24.2%)	3 (15.0%)	1 (9.1%)	0 (0.0%)
Symptomatic case, n (%)	33 (50.8%)	17 (51.5%)	8 (40.0%)	7 (63.6%)	1 (100.0%)
Conjunctival hyperemia, n (%)	17 (26.2%)	9 (27.3%)	6 (30.0%)	2 (18.2%)	0 (0.0%)
Corneal edema, n (%)	30 (46.2%)	15 (45.5%)	9 (45.0%)	5 (45.5%)	1 (100.0%)
Keratic precipitate, n (%)	58 (89.2%)	28 (84.8%)	19 (95.0%)	10 (90.9%)	1 (100.0%)
Vascularization of cornea, n (%)	13 (20.0%)	9 (27.3%)	3 (15.0%)	1 (9.1%)	0 (0.0%)
Peripheral anterior synechia, n (%)	9 (13.8%)	4 (12.1%)	4 (20.0%)	1 (9.1%)	0 (0.0%)
Rejection line, n (%)	4 (6.2%)	4 (12.1%)	0 (0.0%)	0 (0.0%)	0 (0.0%)
Graft outcome, n (%)
[1] Recovered	50 (76.9%)	23 (69.7%)	17 (85.0%)	9 (81.8%)	1 (100.0%)
[2] Failed	15 (23.1%)	10 (30.3%)	3 (15.0%)	2 (18.2%)	0 (0.0%)

DMEK, Descemet’s membrane endothelial keratoplasty; (n)DSAEK, (Non-)Descemet’s stripping automated endothelial keratoplasty; PK, penetrating keratoplasty.

Table 6. Incidence rate of rejection episode

	Surgical procedure of keratoplasty
	[A] PK	[B] DSAEK	[C] nDSAEK	[D] DMEK
Overall summary statistics
Number of patients	198	277	138	117
Person-years [100 person-years]	9.57	8.54	7.09	4.15
Rejection episodes	33	20	11	1
Incidence rate [/100 person-years]	3.45	2.34	1.55	0.24
Incidence rate of rejection episodes by time period after keratoplasty (number of episodes / 100 person-years)
[1] 0-<1 year after keratoplasty	7.55 (14 / 1.85)	4.22 (11 / 2.61)	4.56 (6 / 1.31)	0.00 (0 / 1.16)
[2] 1-<2 years after keratoplasty	6.44 (10 / 1.55)	0.96 (2 / 2.09)	1.83 (2 / 1.09)	0.92 (1 / 1.08)
[3] 2-<3 years after keratoplasty	0.78 (1 / 1.27)	2.83 (4 / 1.42)	2.21 (2 / 0.91)	0.00 (0 / 0.80)
[4] 3-<4 years after keratoplasty	2.93 (3 / 1.03)	3.27 (3 / 0.92)	0.00 (0 / 0.77)	0.00 (0 / 0.46)
[5] 4-<5 years after keratoplasty	1.26 (1 / 0.79)	0.00 (0 / 0.54)	0.00 (0 / 0.70)	0.00 (0 / 0.26)
[6] 5+ years after keratoplasty	1.30 (4 / 3.07)	0.00 (0 / 0.97)	0.44 (1 / 2.30)	0.00 (0 / 0.39)

DMEK, Descemet’s membrane endothelial keratoplasty; (n)DSAEK, (Non-)Descemet’s stripping automated endothelial keratoplasty; PK, penetrating keratoplasty.

Table 7. Univariate and multivariate Cox regression on rejection episode

Characteristic	Univariate model			Multivariate model without eye drop			Multivariate model with eye drop
Characteristic	HR¹	95% CI¹	p-value	HR¹	95% CI¹	p-value	HR¹	95% CI¹	p-value
Surgical procedure			<0.001			0.001			0.003
[A] PK	1.00	—		1.00	—		1.00	—
[B] DSAEK	0.51	0.29, 0.89		0.57	0.32, 1.04		0.55	0.30, 1.01
[C] nDSAEK	0.46	0.23, 0.90		0.56	0.28, 1.14		0.61	0.30, 1.24
[D] DMEK	0.05	0.01, 0.38		0.07	0.01, 0.55		0.08	0.01, 0.61
Age	0.72	0.58, 0.90	0.005	0.77	0.62, 0.96	0.023	0.78	0.62, 0.97	0.032
Sex			0.235
[1] Male	1.00	—
[2] Female	0.74	0.46, 1.21
Hypertension	0.92	0.48, 1.76	0.805
Diabetes mellitus	1.12	0.55, 2.26	0.756
Herpetic keratitis	1.84	0.79, 4.27	0.189
Glaucoma and surgery			0.203
[1] No glaucoma	1.00	—
[2] Glaucoma without surgery	0.55	0.17, 1.77
[3] Glaucoma with surgery	1.44	0.82, 2.51
Prior keratoplasty in opposite eye			0.147
[1] No / Unknown	1.00	—
[2] Yes	1.50	0.88, 2.57
Indication - [01] Failed keratoplasty	1.95	1.17, 3.24	0.014	1.92	1.13, 3.25	0.019	1.73	1.01, 2.96	0.052
Indication - [02] ALI	0.61	0.29, 1.29	0.171
Indication - [03] PBK	0.39	0.12, 1.26	0.068
Indication - [04] Glaucoma surgery	0.89	0.38, 2.06	0.782
Indication - [05] Corneal opacity	1.76	0.84, 3.71	0.163
Indication - [06] FED	0.39	0.10, 1.60	0.127
Indication - [07] XFS	0.49	0.07, 3.57	0.432
Indication - [08] Keratoconus	1.93	0.70, 5.31	0.248
Indication - [09] Perforation	0.85	0.12, 6.11	0.866

Continued onto next page　

Characteristic	Univariate model			Multivariate model without eye drop			Multivariate model with eye drop
Characteristic	HR¹	95% CI¹	p-value	HR¹	95% CI¹	p-value	HR¹	95% CI¹	p-value

Indication - [10] Infection	4.24	1.54, 11.7	0.021	3.60	1.23, 10.5	0.043	2.86	0.96, 8.48	0.092
Indication - [11] CMV corneal endotheliitis	0.00	Not estimable	0.131
Indication - [12] Corneal dystrophy/degeneration	0.94	0.13, 6.81	0.954
Indication - [13] ICE	0.00	Not estimable	0.372
Indication - [14] Others	0.61	0.22, 1.68	0.306
Simple vs. combined surgery			0.433
[1] Simple keratoplasty	1.00	—
[2] Combined keratoplasty	0.76	0.38, 1.54
Graft size			0.277
[1] 4.0-<7.8 mm	1.00	—
[2] 7.6-<8.1 mm	0.58	0.34, 0.99
[3] 8.1-<9.0 mm	0.77	0.29, 2.02
[4] Unknown	0.68	0.21, 2.28
Eye drop at rejection (or last visit)			<0.001						0.019
[1] No	1.00	—					1.00	—
[2] Fluorometholone	1.00	0.46, 2.17					1.37	0.63, 3.00
[3] Betamethasone	2.99	1.34, 6.66					2.68	1.19, 6.04
¹HR = Hazard Ratio, CI = Confidence Interval

Abbreviations: ALI, argon laser iridotomy; CMV, cytomegalovirus; DALK, Deep anterior lamellar keratoplasty; DMEK, Descemet’s membrane endothelial keratoplasty; (n)DSAEK, (non-)Descemet’s stripping automated endothelial keratoplasty; FED, Fuchs’ endothelial dystrophy; ICE, iridocorneal endothelial syndrome; PBK, pseudophakic bullous keratopathy; PK, penetrating keratoplasty; XFS, exfoliation syndrome.

No competing interests reported.

SupplementTableS1S320221021.docx
Supplementary information: The results of subgroup analyses on rejection episodes by demographics, indication for keratoplasty, and simple/combined keratoplasty and graft size that correspond to Tables 1, 2, and 3 are summarized in Supplementary Tables 1 to 3 (Tables S1, S2, S3).

Graft rejection episodes after keratoplasty in Asian eyes

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Discussion

Methods

Declarations

References

Tables

Table 1. Demographics and comorbidities by surgical procedure of keratoplasty

Table 2. Indication of keratoplasty

Table 3. Combined surgery and graft size of keratoplasty

Table 4. Eye drop at rejection (or last visit) and graft outcome after keratoplasty in all cases (N = 730)

Table 5. Findings at rejection episode and graft outcome after the episode [Cases with rejection episode] (N = 65)

Table 6. Incidence rate of rejection episode

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1