Determinants of Epithelial Healing Time and Cumulative Risk-Factor Burden in Eyes with Delayed Corneal Epithelialization After Routine Keratoplasty

Abstract

Background

Delayed epithelialization is a common complication after keratoplasty that can compromise graft survival and increase the economic burden.

Purpose

To identify donor and recipient factors linked to extended epithelial healing time in eyes with delayed epithelialization after routine penetrating and lamellar keratoplasty.

Methods

This retrospective cohort study included eyes undergoing penetrating or lamellar keratoplasty that developed delayed epithelialization, defined as complete epithelial closure occurring ≥8 days postoperatively. Recipient, donor, surgical, and postoperative variables were analyzed. Multivariable linear regression was used to quantify the effect of predictors on epithelial healing time (days). Time to epithelial closure was additionally analyzed using Kaplan–Meier estimates and Cox proportional hazards regression. A cumulative risk-factor burden model was constructed based on independent predictors.

Results

A total of 146 eyes were included. In multivariable analysis, three independent predictors of prolonged epithelial healing were identified: recipient vernal keratoconjunctivitis (VKC) (β = +9.51 days, p < 0.001), female donor sex (β = +2.13 days, p = 0.030), and donor history of chemotherapy or radiotherapy (β = +3.28 days, p = 0.013). Cox regression confirmed slower epithelial healing associated with VKC (HR 0.38), female donor sex (HR 0.65), and donor chemotherapy/radiotherapy exposure (HR 0.64). Kaplan–Meier analysis demonstrated progressively delayed epithelial closure with increasing risk-factor burden (log-rank p < 0.001), with median healing times of approximately 10, 12, and 16 days for patients with 0, 1, and ≥2 risk factors, respectively.

Conclusion

VKC, female donor sex, and donor history of chemotherapy or radiotherapy were independently associated with prolonged epithelial healing time in this single-center cohort of eyes that already exhibited delayed epithelialization after routine keratoplasty. The cumulative risk-factor burden model showed a stepwise association with slower epithelial closure; however, these findings are exploratory and require prospective external validation before integration into donor allocation strategies or routine clinical protocols.

Introduction

Corneal epithelial healing is crucial to graft success soon after keratoplasty. In healthy individuals, the corneal epithelium usually heals within 7 days through three sequential stages: epithelial migration, proliferation, and differentiation to a mature stratified structure.¹ Persistent epithelial defect (PED), defined as a defect persisting for two weeks or more despite medical therapy,² complicates 15–40% of keratoplasty procedures and is associated with increased risk of infection, stromal melting, and graft failure.^3,4

Multiple factors influence epithelial healing after keratoplasty. Known recipient factors include ocular surface disease, limbal stem cell deficiency, neurotrophic conditions such as diabetes mellitus, and inflammatory disorders.^5,6 Corneal denervation after penetrating keratoplasty disrupts neuroanatomic reflexes that regulate tear secretion and blinking, further compromising epithelial healing.⁷ Donor-related factors such as age, gender, death-to-preservation time, and storage duration have shown variable associations with graft outcomes.^8,9 Surgical factors, including tissue manipulation, suturing technique, and trephination size, may also influence epithelialization.¹⁰ Postoperative factors, particularly high-dose topical corticosteroids, have been implicated in delayed healing in some studies,¹⁰ though results remain inconsistent.¹¹

However, delayed epithelial healing is frequently observed even in the absence of clearly identifiable risk factors. Despite extensive research, comprehensive risk-stratification models that include both donor and recipient factors are still limited, especially in routine keratoplasty. Understanding the combined effect of multiple risk factors would support evidence-based preoperative counseling, targeted postoperative care, and better resource allocation. This study aims to identify factors associated with longer epithelial healing time in eyes already experiencing delayed epithelialization after routine keratoplasty and to develop a risk-factor burden model to aid clinical decision-making.

Methods

Study Design and Setting

This retrospective cohort study was carried out at King Khaled Eye Specialist Hospital (KKESH) in Riyadh, Saudi Arabia. It included all patients who underwent penetrating keratoplasty (PKP) or lamellar keratoplasty (LKP) from June 2014 to February 2020. The study protocol followed the tenets of the Declaration of Helsinki and was approved by the institutional review board. (IRB no: 1923-R) Due to the retrospective nature of the study and the use of anonymized patient data, the Institutional Review Board of KKESH waived the requirement for informed consent.

Inclusion and Exclusion Criteria

Inclusion criteria: Eyes were eligible if they underwent LKP or PKP for optical reasons and showed delayed epithelialization. Delayed epithelialization was defined as complete corneal epithelial closure documented on postoperative day 8 or later, equivalent to failure of complete epithelial closure by postoperative day 7. This threshold was selected because routine clinical assessment was performed on postoperative mornings, whereas surgery timing varied during the day, with some procedures performed early in the morning and others later in the day or evening; therefore, a day-based threshold minimized ambiguity related to exact postoperative hours. Other inclusion criteria were complete medical records with documented epithelial healing time and at least 3 months of follow-up.

Exclusion criteria: Eyes that underwent therapeutic or tectonic keratoplasty, documented limbal stem cell deficiency (LSCD), eyes with neurotrophic keratopathy, severe dry eye disease, and severe ocular surface disease were excluded from the study. Additionally, eyes with incomplete medical records and follow-up periods of less than 3 months were excluded.

At our hospital, after routine PKP and LKP, all patients are admitted or monitored daily until the graft is fully epithelialized. They receive topical antibiotics, steroids four times a day, and lubricants. If the epithelium has not healed by day 7, patients are started on autologous serum eye drops and eye patching. If healing has not occurred on day 14, patients undergo an amniotic membrane transplant.

Data Collection

Medical records were reviewed to extract 106 predefined variables that could influence epithelial healing, including the following:

Recipient factors included age and sex, systemic conditions such as diabetes mellitus, connective tissue disease, malignancy, and immunosuppression, as well as ocular conditions like history of vernal keratoconjunctivitis (VKC), dry eyes, use of glaucoma or other eye drops, and mild ocular surface conditions such as old trachoma or allergic eye disease. Indications for keratoplasty were also documented. VKC was recorded when a documented preoperative history of vernal keratoconjunctivitis was present in the medical record. Eyes with documented LSCD, neurotrophic keratopathy, severe dry eye disease, or severe ocular surface disease were excluded. Therefore, included VKC eyes represented eyes with a history of VKC but without documented LSCD or severe ocular surface disease at the time of keratoplasty. In accordance with institutional practice, VKC eyes were clinically quiet and optimized before keratoplasty.

Donor factors included age and sex, systemic conditions such as diabetes mellitus, connective tissue disease, malignancy, and history of chemotherapy or radiotherapy; death-to-preservation time; total preservation time; and donor epithelial sloughing, graded as none, mild, or moderate. Detailed information regarding the timing or type of chemotherapy or radiotherapy was not consistently available in eye bank documentation; therefore, donor cytotoxic exposure was recorded as a binary variable based on available donor medical history.

Surgical factors included the type of keratoplasty (PKP vs LKP), any combined procedures such as cataract extraction or vitrectomy, trephination size, surgeon experience level (consultant vs supervised fellow), and the duration of surgery.

Postoperative factors included the frequency of topical corticosteroids, the type and frequency of lubricants, and any complications occurring in the first 3 months after surgery.

Primary Aim

The study aimed to identify factors associated with prolonged healing in a cohort with delayed healing, rather than to assess the overall risk of delayed healing across all keratoplasty cases. Epithelial healing was evaluated by slit-lamp examination with fluorescein staining. Complete epithelialization was defined as the absence of fluorescein staining across the entire corneal surface. Because this was a retrospective study based on routine clinical documentation, formal interobserver variability testing was not available, and treating surgeons were not prospectively blinded to donor or host characteristics.

Statistical Analysis

Continuous variables were expressed as mean ± standard deviation or median (interquartile range) as appropriate. Categorical variables were expressed as frequencies and percentages. Bivariate analysis was performed using Student’s t-test for continuous variables and chi-square test or Fisher’s exact test for categorical variables.

Multiple linear regression analysis was performed to identify independent predictors of epithelial healing time. Variables with p < 0.10 in bivariate analysis were included in the multivariate model. Cox proportional hazards regression was used to model time to epithelial healing, with hazard ratios (HR) and 95% confidence intervals (CI) reported.

A risk-factor burden model was developed by categorizing patients by the number of independent risk factors present (0, 1, or ≥2). Kaplan-Meier survival analysis was used to estimate median healing times for each risk group, and the log-rank test was used to compare groups.

Statistical significance was set at p < 0.05 (two-tailed). All analyses were performed using SPSS version 25.0 (IBM Corp., Armonk, NY, USA).

Results

Baseline Characteristics

A total of 146 eyes from 146 patients were included in the analysis. The mean age was 42.1 ± 17.3 years, with 82 (56.2%) males. PKP was performed in 91 eyes (62.3%) and LKP in 55 eyes (37.7%). The most common indication was keratoconus/ectasia/dystrophy (88 eyes, 58.9%), and eleven eyes (7.5%) underwent repeat keratoplasty. See Table 1 for more details.

Table 1.

Baseline Characteristics of Eyes with Delayed Epithelialization After Keratoplasty (n = 146)

Variable	Value
Recipient Characteristics
Age, mean ± SD (years)	42.1 ± 17.3
Age >50 years	38 (26.0%)
Male sex	82 (56.2%)
Vernal keratoconjunctivitis (VKC)	28 (19.2%)
Mild ocular surface disease (any)*	54 (37.0%)
Diabetes	21 (14.3%)
Connective Tissue Disease	5 (3.4%)
Any Carcinoma	2 (1.3%)
Taking immunosuppression	7 (4.7%)
Donor Characteristics
Female donor sex	58 (39.7%)
Donor history of radio/chemotherapy	11 (7.5%)
Donor–recipient sex mismatch	63 (43.2%)
Diabetes	50 (34.2%)
Connective Tissue Disease	15 (10.3%)
Any cancer	49 (33.6%)
Preservation time, mean ± SD (Days)	10.73 ± 3.81
Endothelial Cell Count, mean + SD (cells/mm)2	2717 ± 336
Donor Epithelial Sloughing Grade
Mild	119 (81.5%)
Moderate	25 (17.1%)
Severe	2 (1.4%)
Surgical Characteristics
Penetrating keratoplasty (PKP)	91 (62.3%)
Lamellar keratoplasty (LKP)	55 (37.7%)
Combined surgical procedure	19 (13.0%)
Indication for Surgery
Keratoconus/ectasia/ dystrophy	88 (58.9%)
Corneal Scar	22 (15.1%)
Bullous Keratopathy	27 (18.5%)
Re-PKP/LKP (4 & 7 respectively)	11 (7.5%)

Notes: *Mild ocular surface disease include mild dry eyes, allergic eye disease (not Vernal or Atopic keratoconjunctivitis), old trachoma with no sequele mild to moderate meibomian gland dysfunction, presence of pterygium.

Among recipient factors, 28 patients (19.2%) had a history of VKC, 54 (37.0%) had mild ocular surface disease, and 21 (14.3%) had diabetes mellitus. Donor-recipient sex mismatch was present in 63 cases (43.2%). Eleven donors (7.5%) had a history of chemotherapy or radiotherapy for malignancy. See main details in Table 1.

Healing Times

The mean epithelial healing time was 12.9 ± 6.3 days, with a median of 11 days. Among these thirty-two eyes (21.9%) took longer than 14 days to heal.

Bivariate Analysis

In bivariate analysis, the following factors were significantly associated with delayed epithelial healing: recipient history of VKC (p < 0.001), female donor sex (p = 0.010), donor history of chemotherapy/radiotherapy (p = 0.011), and the use of topical corticosteroids >4 times daily (p = 0.049). Donor age, recipient age, mild ocular surface disease, use of glaucoma drops, diabetes mellitus, total preservation time, and donor epithelial sloughing showed no significant association with healing time. A comprehensive summary of bivariate/univariable associations for key evaluated recipient, donor, surgical, and postoperative variables is provided in Supplementary Table 1. In this table, positive β coefficients indicate longer epithelial healing time, whereas Cox HR values below 1 indicate a lower rate of epithelial closure and therefore slower epithelial healing.

Multiple Linear Regression Analysis

Multiple linear regression analysis identified three independent predictors of epithelial healing time (Table 2). The history of VKC was the strongest predictor, adding 9.51 days to healing time (95% CI 5.44–13.59, p < 0.001). In this multivariable model, donor sex (female) remained significant while donor–recipient sex mismatch did not, indicating that donor sex was more informative than sex matching. Female donor sex added 2.13 days (95% CI 0.21–4.05, p = 0.030), and donor history of chemotherapy/radiotherapy added 3.28 days (95% CI 0.71–5.85, p = 0.013). The baseline healing time (intercept) for patients without these risk factors was 11.01 days (95% CI 9.77–12.26, p < 0.001). Patients with all three risk factors had an expected healing time of 25.93 days (11.01 + 9.51 + 2.13 + 3.28), representing a 135% increase from baseline.

Table 2.

Multiple Linear Regression for Epithelial Healing Time (Days)

Predictor	β Coefficient (Days)	SE	95% CI	p
Intercept	11.01	0.63	9.77 to 12.26	<0.001
VKC (present)	9.51	2.06	5.44 to 13.59	<0.001
Female donor sex	2.13	0.97	0.21 to 4.05	0.030
Donor radio-/chemotherapy	3.28	1.30	0.71 to 5.85	0.013

Notes: Model R² = 0.42, adjusted R² = 0.40, F = 34.2, p < 0.001. Patients with all three risk factors had an expected healing time of 25.93 days (11.01 + 9.51 + 2.13 + 3.28), representing a 135% increase from baseline.

Abbreviation: VKC, Vernal keratoconjunctivitis.

Cox Proportional Hazards Analysis

Cox proportional hazards regression confirmed the independent effects of these risk factors on time to epithelial healing (Table 3). The history of VKC was associated with a 62% reduction in the rate of epithelial healing (HR 0.38, 95% CI 0.16–0.87, p = 0.023), indicating that VKC eyes healed at only 38% the rate of non-VKC eyes. Female donor sex reduced the healing by 35% (HR 0.65, 95% CI 0.46–0.92, p = 0.016), and donor chemotherapy/radiotherapy by 36% (HR 0.64, 95% CI 0.40–1.02, p = 0.05). Patients with all three risk factors healed at approximately 16% the rate of those with no risk factors (0.38 × 0.65 × 0.64 = 0.158), corresponding to a substantially prolonged healing course.

Table 3.

Multivariable Cox Proportional Hazards Model for Time to Epithelial Healing

Predictor	HR	95% CI	p
VKC (present)	0.38	0.16 to 0.87	0.023
Female donor sex	0.65	0.46 to 0.92	0.016
Donor radio-/chemotherapy	0.64	0.40 to 1.02	0.050

Notes: Model χ² = 18.7, p < 0.001. Patients with all three risk factors healed at approximately 16% the rate of those with no risk factors calculated as 0.38 × 0.65 × 0.64 = 0.158. HR <1 indicates a lower rate of epithelial closure and, therefore, slower epithelial healing. Hazard ratios are adjusted for the other variables included in the model.

Abbreviations: HR, hazard ratio; VKC, vernal keratoconjunctivitis.

Risk-Factor Burden Model

Kaplan–Meier time-to-event analysis demonstrated progressively delayed epithelial healing with increasing risk-factor burden. A composite risk-factor burden variable was constructed from the three primary predictors identified in multivariable regression (VKC, female donor sex, and donor history of radio- or chemotherapy). Patients were stratified into three groups: 0 risk factors (n=75), 1 risk factor (n=53), and ≥2 risk factors (n=18). Median epithelial healing time increased stepwise across these strata (approximately 10, 12, and 16 days, respectively), with clear separation of the healing curves (log-rank p < 0.001) (Figure 1).

Figure 1.

Kaplan–Meier curves for time to epithelial healing stratified by risk-factor burden (0 factors, n=75; 1 factor, n=53; ≥2 factors, n=18). Increasing burden was associated with progressively slower epithelial healing (log-rank p < 0.001).

The progression from 1 to ≥2 risk factors showed a pronounced effect (+4 days vs +2 days), supporting the concept that cumulative recipient and donor factors may influence the postoperative epithelial healing course. Patients with ≥2 risk factors had a 60% increase in healing time and were significantly more likely to develop a persistent epithelial defect.

Postoperative Complications

As shown in Table 4, postoperative complications related to the graft occurred in 54 patients (37.0%). The most common complications were loose sutures (19 cases, 13.0%), stromal haze (8 cases, 5.5%), and graft failure (7 cases, 4.8%). Eyes that healed >15 days had a higher rate of any postoperative complication compared with those healing within 8–14 days (56.3% vs 31.6%; Fisher’s exact p = 0.013; OR 2.79, 95% CI 1.25–6.21). Severe complications, including corneal melting (3 cases, 2.1%) and need for regraft (7 cases, 4.8%), were associated with prolonged healing time >15 days (Table 4).

Table 4.

Postoperative Complications Stratified by Epithelial Healing Time

Postoperative Complication (n)	8–14 Days (n=114)	≥15 Days (n=32)	OR (≥15 vs 8–14)	95% CI	p (Fisher)
Any postoperative complication (54)	36 (31.6%)	18 (56.2%)	2.79	1.25–6.21	0.013
Corneal melt (3)	1 (0.9%)	2 (6.2%)	7.53	0.66–85.91	0.121
Re-graft (7)	3 (2.6%)	4 (12.5%)	5.29	1.12–24.99	0.041
Haze (8)	5 (4.3%)	3 (9.3%)	2.26	0.51–10.01	0.373
Recurrent epithelial defects (5)	3 (2.6%)	2 (6.2%)	2.47	0.39–15.44	0.301
Re-suturing (19)	12 (10.5%)	7 (21.9%)	2.38	0.85–6.66	0.133
Microbial keratitis (3)	2 (1.8%)	1 (3.1%)	1.81	0.16–20.59	0.527
Graft failure (7)	5 (4.4%)	2 (6.2%)	1.45	0.27–7.87	0.648
Rejection (2)	2 (1.8%)	0 (0.0%)	0.69	0.03–14.79	1.000
Scar (1)	1 (0.9%)	0 (0.0%)	1.16	0.05–29.26	1.000

Notes: Groups: healed within 8–14 days (n=114) vs healed ≥15 days (n=32). Percentages are within-group. Odds ratios (OR) are for >15 vs 8–14. P values from Fisher’s exact test.

Discussion

This study identified three independent risk factors for delayed epithelial healing after keratoplasty: recipient history of VKC, female donor sex, and donor history of chemotherapy/radiotherapy. These factors demonstrated additive effects, with patients having ≥2 risk factors experiencing 60% longer healing times and significantly higher rates of severe complications. To our knowledge, this is the first study to develop a quantitative risk-factor burden model for epithelial healing after routine non-high-risk keratoplasty and the first to identify donor chemotherapy/radiotherapy history as an independent predictor.

The history of VKC emerged as the strongest predictor, increasing healing time by 9.51 days and reducing the healing rate by 62%. This finding is consistent with the known pathophysiology of VKC, which involves chronic ocular surface inflammation, limbal stem cell dysfunction, and basement membrane abnormalities.¹² Although VKC may be relatively uncommon in some general keratoplasty populations, it is clinically relevant in Middle Eastern and Saudi practice, particularly among young patients undergoing keratoplasty for keratoconus or ectatic disease. In the present cohort, VKC was present in 28 eyes (19.2%), supporting its relevance to this regional population. A 2021 study from the Middle East, examining the outcomes of keratoplasties in VKC patients, reported higher-than-expected complications and disease reactivation. Authors noted a trend towards delayed epithelial healing in LKP patients but not in PKP patients.¹³ We, however, did not observe any difference in epithelial healing between PKP and LKP eyes. In our cohort, eyes with documented LSCD, neurotrophic keratopathy, severe dry eye disease, or severe ocular surface disease were excluded; therefore, the VKC variable in our study reflects a history of VKC in eyes considered clinically suitable for routine optical keratoplasty rather than eyes with advanced ocular surface failure. Preoperative optimization of VKC with topical immunomodulators, aggressive lubrication, and careful exclusion or treatment of clinically significant ocular surface disease may improve postoperative outcomes.

Female donor sex was associated with a 2.13-day increase in healing time and 35% reduction in healing rate. This finding provides a novel anterior surface perspective to the ongoing debate surrounding donor-recipient sex incompatibility, which was catalyzed by Hopkinson et al¹⁴ In their national cohort study, Hopkinson demonstrated that male-male matched corneas in keratoconus patients were 30% less likely to fail (p = 0.05) and 20% less likely to reject (p = 0.01) compared with H-Y mismatched corneas, highlighting the potential impact of minor histocompatibility antigens.¹⁴ While recent literature indicates that H-Y mismatch may not significantly compromise long-term endothelial survival in low-risk host beds or modern lamellar keratoplasty,¹⁵ there is some evidence that a donor match may improve outcomes in high-risk keratoplasties.^16,17 Similarly, our findings suggest that donor sex continues to exert some measurable influence on early epithelial healing. The biological mechanisms underlying these sex-specific effects on early epithelial healing remain unclear; however, they may be driven by localized inflammatory cytokine cascades, differential presentation of minor histocompatibility antigens at the ocular surface, or inherent sex-based disparities in basement membrane composition and growth factor expression. However, this finding should be interpreted with caution because the current study was retrospective and not designed to investigate the biological mechanism underlying donor sex effects on epithelial regeneration. Therefore, female donor sex should be regarded as an exploratory signal requiring confirmation in independent cohorts before it is considered for donor-selection or allocation decisions.

Our finding that a donor’s history of chemotherapy or radiotherapy is associated with a 3.28-day increase in epithelial healing time represents, to our knowledge, the first quantitative assessment of this donor-related risk factor in keratoplasty outcomes. Although donor chemotherapy/radiotherapy exposure may be uncommon in unselected keratoplasty populations, it was not uncommon in our delayed epithelialization cohort, occurring in 11 donors (7.5%). This association aligns with established physiological and clinical evidence documenting persistent cellular dysfunction in corneal tissue following cytotoxic exposures. The clinical relevance of donor cytotoxic exposure is most compellingly demonstrated in a retrospective eye-bank analysis of 120 penetrating keratoplasty cases, wherein 5 recipients (4.2%) received corneas from donors who had undergone systemic antimetabolite chemotherapy within 8 weeks of death; all five developed dense central subepithelial opacification post-operatively, and 3 of 5 (60%) required regrafting, with histopathology revealing abnormal basement membrane, basal epithelial apoptosis, and epithelial thinning.¹⁸ The authors concluded that these changes, though nonspecific, showed evidence of poor wound healing, poor-quality epithelium, and chronic disruption of normal epithelial basement membrane histology. This striking outcome rate is substantially higher than typical graft failure rates, suggesting that chemotherapy-induced cellular injury persists in donor tissue even after death, likely reflecting irreversible damage to limbal stem cell niches and residual epithelial progenitors. Other clinical series¹⁹ have reported both transient and permanent LSCD following chemotherapy regimens and after radiotherapy, with some patients demonstrating persistent loss of palisades of Vogt on follow-up. Severe long-term ocular surface morbidity, including LSCD and corneal perforation, has also been documented after high-dose external-beam radiotherapy for periocular or sinonasal tumors.^19–21 Together with evidence that cytotoxic injury induces durable mitochondrial and DNA damage that may impair subsequent epithelial regeneration, these data support our observation and suggest that prior exposure to donor chemotherapy or radiotherapy may merit consideration in donor assessment and postoperative planning. It is worth noting that the timing and type of chemotherapy or radiotherapy exposure were not consistently available; the observed association should be interpreted with caution and requires confirmation in prospective eye-bank studies.

The risk-factor burden model demonstrated a clear dose-response relationship, with median healing times of 10, 12, and 16 days for 0, 1, and ≥2 risk factors, respectively. The accelerated effect observed with multiple risk factors (+4 days from 1 to ≥2 factors vs +2 days from 0 to 1 factor) supports the concept that cumulative recipient and donor factors may influence the postoperative epithelial-healing course. However, the model should be interpreted as exploratory and hypothesis-generating rather than as a validated clinical prediction tool. Prospective external validation is required before this model can be incorporated into routine clinical decision-making, donor allocation, or standardized postoperative protocols.

Our complication analysis showed that 66.7% of corneal melts and 57.1% of regrafts occurred in patients with healing times over 15 days, emphasizing the clinical importance of delayed epithelialization. Persistent epithelial defects after keratoplasty severely impair ocular surface immunity, exposing the underlying stroma to enzymatic breakdown and infection. If these defects are not resolved, they serve as a key trigger for severe, sight-threatening complications, especially progressive stromal keratolysis (corneal melting), infectious keratitis, and subsequent allograft rejection or failure.²²

Topical corticosteroid use >4 times daily was weakly associated with delayed healing in bivariate analysis (p = 0.049) but was not significant in multivariate analysis. This finding is consistent with some previous studies¹⁰ but contrasts with others that found no association.¹¹ The relationship between corticosteroids and epithelial healing likely depends on multiple factors, including formulation (drops vs ointment), specific agent, dosing frequency, and underlying inflammation. In our cohort, high-dose steroids were typically prescribed for increased postoperative inflammation, which itself may be a marker of poor healing. Further prospective studies with standardized steroid protocols are needed to clarify this relationship.

Contrary to some previous reports,⁹ we found no significant association between donor age, death-to-preservation time, or total preservation time and epithelial healing. These findings are consistent with the Cornea Donor Study, which demonstrated no significant effect of donor age (12–75 years) on 5-year graft survival.²³ Similarly, donor epithelial sloughing at the time of surgery did not reach statistical significance in our cohort, contrasting with findings by Feizi et al,²⁴ who reported that donor epithelial sloughing significantly predicted day-1 epithelial defects. However, these studies primarily assessed predictors of delayed epithelialization across full keratoplasty cohorts, whereas our analysis was restricted to otherwise healthy eyes already meeting delayed-healing criteria and modeled the number of days to epithelial closure. This difference in study design may attenuate associations with donor and preservation parameters.

Limitations and Strength

This study has several limitations. The retrospective design allows for missing data and limited documentation of some preoperative risk factors. Because the cohort included only eyes with delayed epithelialization, the analysis identifies associations with healing time among delayed cases rather than population-level predictors of delayed healing. Therefore, the findings should not be generalized to all keratoplasty recipients without validation in broader cohorts. Follow-up duration was variable, with the shortest follow-up being 3 months after surgery. We excluded therapeutic and tectonic grafts to avoid bias from inflamed eyes secondary to infection or perforation, but this limits generalizability to these high-risk populations. The study was conducted at a single tertiary referral center in Saudi Arabia, and findings may not be generalizable to other populations with different prevalence of VKC and other risk factors. Also, due to the retrospective nature of the study, formal interobserver variability testing for epithelial closure assessment was not available, and fluorescein staining was performed as part of routine clinical care rather than as part of a prospective standardized grading protocol. Important ocular-surface variables, including tear-film quality, corneal sensitivity, blink dynamics, meibomian gland status, and detailed inflammatory activity, were not consistently available. Donor chemotherapy/radiotherapy exposure was recorded as a binary variable because the timing, type, dose, and interval from treatment to death were not consistently available in eye-bank documentation.

Conversely, this study assesses a relatively large, well-defined group of low-risk routine keratoplasty eyes with delayed epithelialization, including detailed recipient and donor variables. We employed a time-to-healing approach (Kaplan–Meier/Cox) combined with a straightforward risk-factor burden exploratory model to enhance understanding by clearly demonstrating a dose–response relationship in healing rates.

Future Directions

Future research should focus on prospective validation of the risk-factor burden model in independent cohorts, investigation of biological mechanisms underlying gender-specific effects on epithelial healing, development of standardized protocols for preoperative optimization of high-risk patients, and randomized controlled trials comparing biologic agents (platelet-rich plasma, umbilical cord serum, cenegermin) for prevention and treatment of delayed epithelialization in high-risk patients. Additionally, molecular studies examining the effects of chemotherapy and radiotherapy on corneal tissue quality may inform donor selection criteria and tissue processing protocols.

Conclusion

In this single-center cohort of eyes with delayed epithelialization after routine keratoplasty, recipient VKC, female donor sex, and donor history of chemotherapy/radiotherapy were independently associated with prolonged epithelial healing time. A cumulative risk-factor burden model demonstrated a stepwise association with slower epithelial closure and a higher rate of complications. These findings may support hypothesis generation and closer postoperative awareness in selected delayed-healing eyes; however, prospective external validation is required before integration into donor allocation strategies or routine clinical protocols.

Funding Statement

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data Sharing Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request, subject to institutional review board approval and patient privacy regulations.

Disclosure

The authors declare no conflicts of interest.