Risk Factors for Graft Failure After Penetrating Keratoplasty in Eastern China from 2018 to 2021

Introduction

Corneal disease is one of the leading causes of blindness or severe vision impairment in the global population (especially in developing countries in Asia and Africa), and the condition affects millions of people [1,2]. The main method for restoring vision to patients with corneal blindness is corneal transplantation, which mainly includes penetrating keratoplasty (PK), lamellar keratoplasty, and endothelial keratoplasty [1]. Although lamellar keratoplasty has been widely promoted worldwide in the past decade, PKs are still the main procedure, accounting for more than 50% of corneal transplantations in China [3,4]. Compared with other transplantations, PKs have a higher rate of graft failure, with a shorter median predicted survival time [5]. Therefore, the determination of the risk factors for graft failures in PKs is an important aspect of preventing and controlling corneal blindness.

Graft failure in PK is usually defined as an irreversible change in the graft that prevents vision recovery and is mostly attributed to immunological rejection, late endothelial failure (LEF), ocular surface disorders, postoperative glaucoma, and wound dehiscence [6]. To the best of our knowledge, although there have been some previous reports on the factors affecting the success rate of PKs after surgery, analyses of data clarifying the situation in Shanghai (and even in Eastern China) in the past 5 years still need to be established. Thus, in the present study, we reported on the survival of grafts in patients who had undergone PKs in one of the largest keratoplasty centers in Eastern China during the past 4 years. In addition, we strived to analyze the potential factors associated with early, late, or all graft failures.

Material and Methods

STUDY DESIGN:

This retrospective, single-center study was approved by the Ethics Committee of the Eye and ENT Hospital of Fudan University and was conducted in accordance with the Declaration of Helsinki.

Patients who underwent PK at the Eye and ENT Hospital of Fudan University from January 2018 to December 2021 and who were followed up for at least 1 year were included in the retrospective cohort. Demographic, surgical, and follow-up data were collected from the patients’ medical records, including age, sex, marriage, nationality, occupation, education background, household registration, indication, histories of ocular and systematic diseases, surgeon, graft size, techniques of anesthesia and anterior chamber formation, surgical complications (including anterior chamber bleeding, vitreous prolapse, and iris adhesions), the number of stitches, and the intraocular pressure (IOP) after surgeries. The exclusion criteria were as follows: (1) any history of keratoplasty in the study eye; (2) incomplete surgical records or donor information; and (3) no photographic records available during follow-up visits.

DONOR STORAGE AND ENDOTHELIAL CELL ANALYSIS:

All of the donor corneas originated from the Acceptance Station of the Eye and ENT Hospital of Fudan University, Shanghai Red Cross Eye Bank. The corneas were stored in intermediate-term storage media (Bausch & Lomb Inc., Bridgewater, NJ, USA) at 4°C for no more than 14 days. The age of the donor, the central endothelial cell density (ECD), and the preservation time between harvesting and surgery were collected from the eye bank database.

SURGICAL TECHNIQUE:

PKs were performed by corneal surgeons with individual surgical experiences of at least 10 years. The donor buttons were punched out of the endothelial side and sutured to the recipient bed with interrupted 10-0 nylon sutures. Additional surgeries, such as lensectomy, vitrectomy, intraocular lens removal, or implantation, were performed if needed. After the surgery, tobramycin and dexamethasone eye drops were routinely used. Sutures were removed 3 to 24 months after surgery.

GRAFT FAILURE:

Graft failure is defined as (1) clouding of the corneal grafts after surgery that is irreversible under treatment; and (2) reimplantation being required for any reason. The classification of graft failure was the same as that used in previous studies [7], with early graft failure defined as failures presenting between the first postoperative day and 8 weeks and unable to be reversed during 8 weeks, and late graft failure defined as failures presenting between 8 weeks and the end of the 1-year follow-up period. Survival time was calculated as the interval between the date of surgery and the date of the first appearance of corneal opacity (definitive failure) or the last visit (censored or lost-to-follow-up cases).

STATISTICAL ANALYSIS:

Kaplan-Meier survival analysis was used to calculate the cumulative survival rate at 1 year, along with 95% confidence intervals (CIs) and hazard ratio (HR). Cox proportional hazards regression models were used to assess the association of factors with all of the graft failures and late failures. First, we used a univariable base model to select factors with P<0.10. Subsequently, the factors were included in a multivariable backward model selection procedure to build the final multivariable model. Among the selected factors, candidates with P<0.05 were considered to be statistically significant. Logistic regression models were used to assess the association of factors with early failures. The method of variable filtering is the same as mentioned above. In all of the models, the surgeon was considered as a random effect.

In the description of the baseline information, continuous data are expressed as mean±standard error, and the categorical data are shown as percentages. All of the analyses were performed using IBM SPSS Statistics (version 20.0, International Business Machines Corporation, USA).

Results

DEMOGRAPHIC INFORMATION:

The demographic features of the 502 eyes in 481 recipients are summarized in Table 1. A total of 293 recipients (60.91%) were male, and 188 recipients (39.09%) were female. The average follow-up time period was 439 days. Moreover, the average age of the patients at the time of surgery was 40.07±26.20 years, ranging from 1 month to 86 years. The most common indication was corneal opacity, which accounted for 39.24% (197/502) of the study eyes. The graft survival curves for patients with different indications are shown in Figure 1.

ALL FAILURE AND 1-YEAR SURVIVAL RATES:

Rejections occurred in 93 eyes, of which 36 eyes (38.71%) were classified as early failure and 57 eyes (61.29%) as late failure. The overall failure rate was 18.53% (93/502).

The multivariate analysis that was performed on the overall data and the 1-year cumulative graft survival rate are shown in Table 2. Predictors with corresponding P values <0.10 in the univariate base model were included in the subsequent multivariate model for selection. In the final multivariate model, factors independently associated with an increased risk of graft failure included burns (HR: 12.56, 95% CI [2.4, 65.71]), nonsuppurative inflammation (HR: 11.06, 95% CI [2.49, 49.03]), suppurative inflammation (HR: 12.49, 95% CI [2.76, 56.37]); immune disease (including Sjögren’s syndrome, systemic lupus erythematosus, rheumatoid arthritis, ankylosing spondylitis, and pemphigoid; HR: 1.70, 95% CI [0.64, 4.48]); donor ECD (HR: 0.92/100 cells, 95% CI [0.87, 0.98]); and surgical complications (HR: 1.83, 95% CI [1.06, 3.17]). The 1-year survival rates of patients with these factors were clearly lower than the survival rates of patients without these factors.

Recipient sex, age, occupation, completion of compulsory education, marriage, diabetes history, cataract history, anterior chamber formation during surgery, donor age, cornea preservation time, and IOP after surgery were included in the base model but not in the final model.

EARLY FAILURES:

Factors related to graft failures within 8 weeks, in the base and final models, are displayed in Table 3. Factors independently associated with an increased risk of early failure included burns (OR: 18.11, 95% CI [1.70–192.37]), nonsuppurative inflammation (OR: 12.07, 95% CI [1.49–97.75]), and suppurative inflammation (OR: 13.56, 95% CI [1.69–109.14]); graft size (OR: 2.40, 95% CI [1.28–4.45]); and ECD (OR: 0.85/100 cells, 95% CI [0.75–0.95]).

The recipient’s age, cataract history, and IOP after surgery were included in the base model but not in the final model.

LATE FAILURES:

Factors related to graft failures from the 9th week to 1 year in the base and final models are displayed in Table 4. We found that in addition to the physical factors identified (nonsuppurative inflammation [HR: 5.95, 95% CI {1.28–27.64}], suppurative inflammation [HR: 6.36, 95% CI {1.27–31.75}], surgical complications [HR: 1.99, 95% CI {1.01–4.00}], donor ECD [HR: 0.93/100 cells, 95% CI {0.86–1.2}], and donor age [HR: 1.12/10 years, 95% CI {0.86–1.46}]), 2 social factors were negatively associated with late failures: completion of compulsory education (HR: 0.45, 95% CI [0.21–0.94]) and marriage (HR: 0.27, 95% CI [0.09–0.81]).

The recipient sex, age, occupation, history of diabetes, cataract and immune diseases, anterior chamber formation, the number of stitches in surgery, cornea preservation time, and IOP after surgery were included in the base model but not in the final model.

Discussion

RECIPIENT FACTORS: INDICATIONS:

The information that we obtained indicated that corneal opacity, septic and nonseptic keratitis, and keratoconus were the main indications for PK in our center, which is consistent with the summary of indications for corneal transplantation throughout China during 2014–2018, as reported by Gao et al [4]. Previous literature has reported that preoperative diagnosis has a significant influence on graft survival. Keratoconus showed the highest 1-year survival rate (98.2%) in our data, and was considered to have the best prognosis, which is consistent with existing studies [10–12]. Moreover, keratoplasties in eyes with active inflammation caused by trauma, burn, rejection, or infection (also known as high-risk transplantations) have a much higher incidence of immune-mediated graft rejection [13]. This is mainly attributed to the stimulation of lymphatic vessels and angiogenesis by the inflammatory environment, through which alloantigen efflux and effector T-cell recruitment can occur [14].

The decompensation of the corneal endothelium is another important reason for graft failure. Xu et al [15] and Chung et al [16] reported that patients with corneal endothelial loss had a faster rate of endothelial cell loss (ECL), especially at 3 months postsurgery. This may be related to the migration of donor endothelial cells to the peripheral area of the recipient cornea. Due to the fact that only a small percentage of our patients had their ECD values examined in follow-up visits, we did not validate this opinion.

RECIPIENT FACTORS: IMMUNE DISEASE:

Immunological rejection accounts for approximately 30% of corneal graft failures, in which the host’s immune system recognizes the allograft and produces an immune response [6,17]. Patients with autoimmune or allergic diseases usually have an active innate and adaptive immune system that can produce higher levels of inflammatory cytokines and antibodies that may be a risk factor for graft rejection [18,19]. However, the association between immune disorders and corneal graft failure is less commonly reported in previous reports. For example, Jhanji et al reported a case of presumed graft rejection after DALK in a patient with lupus erythematosus after stopping topical corticosteroids in 2010, and they proposed that the presence of more reactive antibodies may be the reason for the higher incidence of rejection after organ transplantation in patients with systemic lupus erythematosus [20]. In recent years, the widespread use of immune checkpoint inhibitors in cancer treatment has also caused ophthalmologists to notice the relationship between an overactive immune response and transplant rejection [21]. However, only 12 patients with immune disease were included in our cohort, which made analysis at the subgroup level impossible.

SURGICAL FACTORS: SURGICAL COMPLICATIONS:

The relationship between surgical complications and graft failure has been mentioned in similar studies, with most scholars agreeing that surgical complications are a risk factor for graft failure [7]. Cruz et al evaluated potential risk factors for surgical complications in 258 PK recipients and showed that a history of ocular surgery, combined cataract extraction, and >8 mm implant beds were associated with a higher risk of surgical complications [22]. However, we did not observe a direct relationship between the first 2 factors and transplant failure. Positive vitreous pressure caused by anterior chamber exposure is the potential reason why PK has a higher risk of suprachoroidal hemorrhage and vitreous prolapse compared with DALK [23,24]. However, further analysis needs to be supported by more detailed surgical records.

SURGICAL FACTORS: GRAFT SIZE:

The mean diameter of the grafts in our data was 7.56±0.69 mm, ranging from 5 mm to 9.5 mm. Our observation that a larger implant size is a risk factor for early graft rejection is consistent with the classical results of Vail et al [25]. Li et al and Alfaro Rangel et al reported that PKs with grafts ranging from 8.75–10.0 mm in diameter (also known as large diameter penetrating keratoplasty [LDPK]) are often used in severe infectious keratitis [26,27]. Thus, the observed poor prognosis is also related to the consequences of the disease itself, including the recurrence of infection, long-term inflammatory disruption of anterior chamber-associated immune deviation, and secondary glaucoma. A retrospective study on patients receiving LDPK conducted by Skeens et al reported that when combined with proper local anti-inflammatory therapy (including topical steroids and cyclosporine), LDPK can result in less postoperative astigmatism than procedures using small-diameter grafts, without an increased risk of graft failure [28]. However, they still mentioned that proximity to the limbus may be a potential risk factor.

DONOR FACTORS:

The only donor factors that we observed to have an effect on graft failure were donor age and ECD. The effect of donor age on late failure is consistent with that observed in the Cornea Donor Study (CDS), which reported lower success rates in donors aged 72–75 years than in younger donors but was only demonstrated at the long-term follow-up (6 years) [29]. In addition, CDS also showed that, compared with corneas from donors aged less than 65 years old, those from donors aged greater than 65 years had a faster rate of ECD decline and a higher risk of endothelial decompensation after surgery [30]. In our data, the rate of graft failure within 1 year was significantly higher when the donors were older than 71 years than when the donors were younger than 71 years, thus suggesting that clinicians may need to pay earlier attention to corneas from donors of extreme ages. Schön et al reported that risk factors associated with low ECD included age, sex, death-to-explantation interval (DEI), and pseudophakic eyes [31]. To maximize the collection of available corneas, they presented a model that combines these 4 factors to predict ECD, which we believe will greatly assist in donor screening.

SOCIAL FACTORS:

We found that, among the recipients, 2 factors were protective against late failure: marriage and the completion of China’s 9-year compulsory education. In contrast, the recipients’ household registration, occupation, and status of whether or not they lived in a rural area were not associated with any failure. We believe that education and the care of family members are associated with a higher quality of life and better hygiene practices, which are beneficial for rehabilitation after corneal transplant surgery. Patients with better living conditions also tend to seek medical care earlier after experiencing discomfort, which is thought to be associated with better drug efficacy and higher postoperative vision [32]. Therefore, we hope that the government and society can realize the importance of health education in the prevention and treatment of eye diseases and strengthen health care for people who are living alone to prevent the occurrence of corneal blindness.

OTHER FACTORS THAT DIFFER BETWEEN EARLY FAILURE, LATE FAILURE, AND ALL FAILURE:

In many previous articles on other types of corneal transplantation, it has been mentioned that the factors leading to early and late transplant failures are different. Wilson et al highlighted in their study that primary disease and corneal endothelial cell count are significant factors influencing the whole period of corneal graft survival [33]. Rafael et al further confirmed that corneal burns and infectious keratitis are important contributors to both early and late corneal graft rejection, which aligns with our research findings [34]. Terry et al found in their study that surgical complications and donor history of diabetes mellitus are crucial factors leading to early rejection, which correlates with our findings on the role of surgical factors in early corneal graft rejection [35]. Terry et al also identified primary disease as the sole factor influencing late rejection [35]. While partially consistent with our results, our study emphasizes that corneal graft survival depends not only on medical efforts but also on patient care in daily life. Therefore, we integrated various social factors into our analysis and found that higher education levels and more familial support contribute significantly to corneal graft survival. This finding represents a novel contribution to the field. In conclusion, we believe that in addition to ensuring optimal surgical procedures and donor materials, providing patients with extensive postoperative education, improving their medical compliance, and offering adequate emotional support can significantly enhance the long-term survival of corneal grafts.

LIMITATIONS:

There were several limitations of this study. First, this was a retrospective single-center study, and the data records were not all completed and standardized. For example, detailed preoperative diagnoses in some recipients, the donors’ medical histories, and the causes of death were not available for researchers. Some patients were not followed up regularly after surgery, and parameters such as IOP and visual acuity were not continuously monitored after surgery. Second, the available medical records only support the determination of whether failure occurred by photographs and admission records and cannot accurately classify the causes of graft failure. In addition, the short follow-up time period meant that our data cannot verify the pathological differences between earlier and later graft failures. For example, Kelly et al proposed that failure within the first 10 years after surgery was mainly related to rejective inflammation, and failure occurring later may be caused by noninflammatory diseases [12]. Therefore, long-term prospective studies and follow-up are needed to further understand the potential risk factors for transplant failure.

Conclusions

In conclusion, we found that burns and suppurative or nonsuppurative inflammation as indications, immune disease of the recipient, surgical complications, and lower endothelial cell counts in cornea donors were predictors of a higher risk of failure. In addition, social factors, including education and marriage, also play a role in the health of patients with corneal diseases.