Clinical Trials in Cornea

All content on Eyewiki is protected by copyright law and the Terms of Service. This content may not be reproduced, copied, or put into any artificial intelligence program, including large language and generative AI models, without permission from the Academy.

Article initiated by:

Francesc March de Ribot, MD, PhD

All contributors:

Amber Yuna Kwon, BA, Apoorva Chowdhary, Jainam Shah, Olivia Lee, MD, TenayaAnn Tresko, BS, Jessica Tabary, Vatinee Y. Bunya, MD, MSCE, Francesc March de Ribot, MD, PhD, Shu Feng, MD, Minh T Nguyen, MD, Colleen Halfpenny, M.D., Brian Chou, MD, Megan Kasetty, MD, Sarah Zhou, MD, Dina Abdelsalam

Assigned editor:

Review:

Assigned status Update Pending

.

add
Contributing Editors:	add

Herpetic Eye Disease Study - 1 (HEDS 1)

Ophthalmology 1994;1871-1882| Ophthalmology 1994;1883-1896| Arch Ophthalmol 1996;1065-1072.

Objectives

The goal was to determine the role of 1) topical steroids in stromal keratitis associated with HSV; and 2) oral acyclovir in HSV stromal keratitis and HSV iridocyclitis in eyes being treated with topical trifluridine and topical steroid.

Design

3 separate randomized, double-masked, placebo-controlled trials to evaluate HSV 1) Stromal keratitis with steroids, 2) without steroids, and 3) Iridocyclitis. Groups:

1. SKN – Stromal keratitis, not on steroids: 106 patients who had not used a topical steroid in the preceding ten days were randomized to receive topical prednisolone or placebo. Topical prednisolone phosphate 1% was initially started at 8x daily for the first week, then tapered over the course of 10 weeks.
2. SKS – Stromal keratitis, on steroids: 104 patients who were already using topical steroids were randomised to receive acyclovir 400 mg five times daily for ten weeks or placebo. Topical steroid and antiviral therapy were standardized for both groups.
3. IRT – Iridocyclitis, receiving topical steroids: 50 patients who were already using topical steroids were randomized to receive acyclovir 400 mg five times daily for ten weeks or placebo. Topical steroid and antiviral therapy were standardized for both groups.

Patients were evaluated weekly for ten weeks, fortnightly for six more weeks, and at six months.

Main outcome measures

Visual acuity, resolution of active disease, treatment failure (defined as worsening/no improvement of stromal keratitis or adverse event).

Results

1. SKN – Stromal keratitis, not on steroids: 106 patients who had not used topical steroid in the preceding ten days were randomized to receive topical prednisolone or placebo. Corticosteroid therapy had a longer time to treatment failure, reduced the risk of persistent or progressive stromal keratouveitis by 68%, and had a faster resolution time. At 6 months, there were no differences in VA or herpetic recurrence between the groups.
2. SKS – Stromal keratitis, on steroids: 104 patients who were already using topical steroids were randomized to receive acyclovir 400 mg five times daily for ten weeks or placebo. There was no difference in treatment failure (75% in acyclovir and 74% in placebo) or worsening of disease (18% in acyclovir and 19% in placebo); however, at 6 months, a greater improvement in VA was noted in the acyclovir group.
3. IRT – Iridocyclitis, receiving topical steroids: 50 patients using topical steroids were randomized to receive acyclovir or placebo. Oral acyclovir added to topical trifluridine and prednisolone resulted in fewer treatment failures in the first ten weeks (50% compared to 68% in the placebo group). Relapse rates were similar between the two groups. Overall, while the clinical trend favored adding oral acyclovir, conclusions lacked statistical significance due to the small number of patients in this arm of the trial.

Limitations

Many patients were lost to follow-up. Current treatment patterns favor topical acyclovir or ganciclovir over trifluridine.

Conclusions

In patients with HSV stromal keratits, the addition of topical corticosteroid significantly reduced persistence or progression of stromal inflammation and shortened time to resolution. The addition of oral acyclovir did not significantly improve treatment success or speed up recovery when patients were already receiving topical steroids and topical antivirals. There was no statistically significant benefit of oral acyclovir in HSV iridocyclitis due to small sample size.

Pearls for clinical practice

Topical steroids can significantly reduce stromal inflammation and duration of keratits in patients with HSV stromal keratitis.

Zoster Eye Disease Study (ZEDS)

Objectives

To determine whether 12 months of suppressive oral valacyclovir compared to placebo reduces the occurrence of herpes zoster ophthalmicus (HZO) related ocular complications.

Design

Multi-site, double-masked, placebo-controlled, randomized clinical trial from November 2017 to June 2024, which included 527 participants age ≥18 years old with a history of an HZO rash and active keratitis or iritis within the past 1 year. All patients were immunocompetent, nonpregnant and had normal kidney function.

Participants were randomized 1:1 to either 1000mg oral valacyclovir or placebo once daily for 12 months in addition to routine care. Follow-up was every 3 months for 18 months total. Participants were also stratified based on age at onset of HZO (<60 or 60+ years) and disease duration at enrollment (<6 or 6+ months).

Main outcome measures

Time to first occurrence of worsening stromal keratitis (SK), endothelial keratitis (EK), iritis, or dendritiform epithelial keratitis (DEK) at 12 months, and a secondary endpoint at 18 months.

Results

Role of suppressive valacyclovir for the prevention of new or worsening HZO ocular disease

While the primary endpoint did not reach statistical significance at 12 months, long-term suppressive valacyclovir demonstrated a cumulative treatment benefit by 18 months in reducing the risk of new or worsening corneal complications such as stromal keratitis, endothelial keratitis, iritis, or dendriform epithelial keratitis.

At 12 months (primary endpoint), there was no significant reduction in new or worsening ocular disease between the valacyclovir group and the placebo group (28% versus 33%; HR 0.77; 95% CI 0.56–1.05; P=0.09). However, by 18 months (secondary endpoint: 6 months of post-treatment observation), the valacyclovir group showed a statistically significant reduction in these events (32% in valacyclovir group versus 40% in the placebo group; HR 0.73; 95% CI 0.55–0.97; P=0.03). Subgroup analysis suggested a more pronounced benefit in younger participants with recent disease onset at 12 months, and a benefit for all recent-onset participants at 12 and 18 months.

Multiple/recurrent episodes of keratitis or iritis

Valacyclovir significantly reduced the frequency of multiple/recurrent episodes of HZO-related keratitis and iritis throughout the study period. Unlike the primary endpoint (which only measured the time to the first event), the analysis of multiple recurrences showed a clear benefit for suppressive therapy. Valacyclovir treatment resulted in a significantly lower risk of experiencing repeat episodes at both 12 and 18 months.

Treatment Safety and Tolerability

Amongst the study participants, the usage of oral valacyclovir 1000mg once daily during the study period had an excellent safety profile for long-term valacyclovir use. There were no recorded episodes of acute kidney failure or suspected unexpected serious adverse reactions related to the treatment.

Limitations

• Enrollment fell below expectations resulting in reduced power of the study.
• The ideal dose and duration of antiviral suppressive treatment was not tested.
• The study results are specific to immunocompetent individuals and those with normal kidney function.

Conclusions

Patients treated with prophylactic valacyclovir 1000mg once daily for 12 months had significantly fewer recurring episodes of keratitis or iritis at 12 and 18 months.

Clinical Pearls

Consider at least a 12-month course of suppressive valacyclovir for patients with HZO-related ocular complications, especially in those with recent-onset disease or history of recurrent keratitis or iritis.

The Steroids for Corneal Ulcers Trial (SCUT)

Objective

To assess whether use of topical corticosteroid as adjunctive treatment for bacterial corneal ulcers has a significant effect on clinical outcomes.

Design

Multi-center, placebo-controlled, randomized, double-masked clinical trial, from September 1, 2006 to February 22, 2010. Of the 1769 patients screened, 500 enrolled, with 442 and 399 returning for their 3- and 12-month follow ups, respectively.

Exclusion criteria included evidence of fungal, acanthamoeba, or herpetic infection, corneal perforation or impending perforation, prior penetrating keratoplasty, use of topical or systemic steroid during present clinical course, and vision less than 6/60.

After confirming diagnosis of a bacterial cornea ulcer via positive culture, participants received topical 0.5% moxifloxacin dosed at 1 drop every hour while awake for 48 hours. Patients were then randomly divided into one of two groups:

• Treatment group (n=250)
  • Patients were treated with topical 1.0% prednisolone sodium phosphate dosed at 1 drop 4 times per day for 1 week, then twice a day for 1 week, and then once a day for 1 week.
• Placebo group (n=250)
  • Patients were treated with topical 0.9% sodium chloride dosed at 1 drop 4 times per day for 1 week, then twice a day for 1 week, and then once a day for 1 week.

Both groups continued to receive topical 1.0% moxifloxacin after the initial 48 hours, dosed at 1 drop applied every 2 hours until reepithelialization, then 4 times a day until 3 weeks from enrollment.

Main Outcome Measures

Improvement in best corrected visual acuity (BCVA) at 3 and 12 months from enrollment, defined as an improvement of two or more lines. Secondary outcomes included infiltrate or scar size, time to reepithelialization, and adverse events such as corneal perforation, elevation in intraocular pressure (IOP), or a worsening corneal ulcer.

Results

There was no significant difference in 3-month BCVA, infiltrate or scar size, reepithelialization time, corneal perforation or worsening ulcer events between the corticosteroid and placebo groups.

In a 3-month subgroup analyses, use of corticosteroids showed significant effect depending on baseline BCVA, ulcer location, and infiltrate depth:

• Patients with baseline BCVA of counting fingers or worse who received topical corticosteroid had approximately 1.7 lines better BCVA compared to the placebo group at 3 months.
• Patients with ulcers covering the central pupil (4 mm) who received topical corticosteroid had approximately 2 lines better BCVA compared to the placebo group at 3 months.
• Patients with greatest ulcer infiltrate depth (67%-100%) who received topical corticosteroid had approximately 1.5 lines better BCVA compared to the placebo group at 3 months.

Additionally, at 3 months there was a statistically significant greater number of patients in the placebo group who developed an IOP greater than 25 mmHg but less than 35 mmHg. There were no IOP elevations above 35 mmHg in either the treatment or placebo group.

At 12 months, there was no significant difference in BCVA or scar size between the treatment and placebo group. A subgroup analysis of Nocardia vs. non-Nocardia ulcers showed a statistically significant 1-line improvement in BCVA among non-Nocardia ulcers that received topical corticosteroid; there was no significant improvement in BCVA for Nocardia-positive ulcers. Additionally, Nocardia-positive ulcers displayed a statistically significant larger scar size (0.47mm larger) with use of topical corticosteroid; there was no significant difference in scar size for non-Nocardia ulcers.

The most common organism cultured was Streptococcus pneumoniae, followed by Pseudomonas aeruginosa, Nocardia species, Staphylococcus species, Moraxella species, Streptococcus viridans group, Corynebacterium species, Klebsiella species = Pseudomonas species (non-aeruginosa), Enterobacter species, and Bacillus species = Mycobacteria species = Haemophilus influenzae.

Limitations

• Most enrollees were non-contact lens wearing individuals in India.
• Fungal, acanthamoeba, or herpetic ulcers were not included.
• To standardize the antibiotic that was used (a broad-spectrum fluoroquinolone), some patients may not have been treated with the most efficacious antibiotic for their specific needs, given the wide range of organisms that were cultured.
• Physicians who thought an alternative antibiotic treatment was medically necessary could change or add antibiotics at any time; the rate of antibiotic change was approximately 15% in both the treatment and placebo groups.
• Nocardia species was the 3^rd most isolated bacterial organism in the trial, which is not a commonly reported organism in the United States and Europe.

Conclusions

Use of topical corticosteroid as adjunctive treatment for bacterial corneal ulcers does not improve vision at 3 and 12 months compared to placebo. A subgroup of patients at 3 months with particularly severe ulcers at baseline had a significant improvement in visual acuity following topical corticosteroid use. A subgroup of patients at 12 months with non-Nocardia ulcers had a significant improvement in visual acuity following topical corticosteroid use, and Nocardia-positive ulcers displayed a significantly larger scar size following topical corticosteroid use. Use of corticosteroids was not associated with an increase in IOP at 3 months. There was no significant difference in adverse events such as corneal perforation with use of topical corticosteroids.

Clinical Pearls

• Overall, topical corticosteroids in the treatment of mild to moderate bacterial corneal ulcers do not have significant effect on clinical outcomes at 3 and 12 months.
• However, patients with the greatest ulcer severity (i.e. worst visual acuity or central ulcer location at baseline) show greater improvement in vision at 3 months after using topical corticosteroids.
• Corneal ulcers that are culture-negative for Nocardia species are associated with greater improvement to vision at 12 months after using topical corticosteroids.

Steroids and Cross-Linking for Ulcer Treatment (SCUT II)

Objective

To determine whether adjunctive corneal cross-linking (CXL) and early topical corticosteroids improve clinical outcomes in bacterial keratitis compared with antibiotic therapy alone.

Design

Multi-center, sham and placebo-controlled, randomized, double-masked three-arm clinical trial, from September 2020 to October 2023 at Aravind Eye Hospitals in India and Bascom Palmar Eye Institute, University of Miami, FL, USA. Of the 1992 patients with smear-positive bacterial corneal ulcers screened, 280 participants (14%) were enrolled (mean age 51 years; 65% male), with 207 and 219 returning for their 3- and 6-month follow ups, respectively.

Exclusion criteria included evidence of concomitant infection on exam, gram stain, or confocal microscopy, impeding or current perforation, scleral involvement at presentation, history of corneal transplantation or intraocular surgery within the past three months, and pinhole visual acuity worse than 20/200 in the unaffected eye.

After confirming diagnosis of a bacterial cornea ulcer via positive culture, participants received a 30-minute loading dose of topical 0.1% riboflavin and 20% dextran T500 drops every 2 minutes. Patients were then randomly divided into one of three groups:

• Group 1, Standard Therapy Group (n=93)
  • Patients were treated with topical 0.5% moxifloxacin plus the topical placebo and sham UVX therapy.
• Group 2, Early Steroids (n=93)
  • Patients were treated with  topical 0.5% moxifloxacin plus topical difluprednate 0.05% and sham UVX therapy.
• Group 3, Cross-Linking (n=93)
  • Patients were treated with topical 0.5% moxifloxacin and topical difluprednate 0.05% plus corneal cross-linking. Cross-linking was performed with exposure to UV-A light at a wavelength of 365 nm with an irradiance of 3 mW/cm² for 30 minutes, or a total dose of 5.4 J/cm². The UV lamp used depended on the county; in India the PESCHKE Meditrade GmbH (Hueneberg, Switzerland) was used, while in the United States the Avedro KXL System (Waltham, MA, USA) was used.

In all study participants, 0.5% moxifloxacin was given every hour while awake for the first 48 hours, then extended to every 2 hours while awake until re-epithelization. Topical placebo or corticosteroid drop was started 24 hours after the initiation of antibiotics and dosed at 1 drop 4 times per day for 1 week, then three times a day for 1 week, followed by two times a day for 1 week, and then once a day for 1 week.

Main Outcome Measures

Improvement in best corrected visual acuity (BCVA) at 6 months from enrollment. Secondary outcomes included BCVA at 3 weeks and 3 months, infiltrate and/or scar size, subjective pain level, microbiological cure at 2 days, and adverse events such as corneal perforation or endophthalmitis.

Results

There was no significant benefit in 6-month BCVA between adjunctive steroids and placebo, or between adjunctive CXL plus steroids and steroids alone. There was no significant difference in re-epithelialization time, corneal perforation or worsening ulcer events across groups.

In 6-month analyses, neither adjunctive steroids nor CXL plus steroids improved outcomes when compared with topical antibiotic monotherapy, and CXL was associated with greater scarring and pain:

• Scar and infiltrate size did not differ significantly between steroids alone and placebo; adding CXL to steroids increased scar size compared with steroids alone (0.56 mm; 95% CI, 0.20-0.92; P=0.02).
• Pain scores did not differ significantly between steroids and placebo but were significantly higher with CXL plus steroids than with steroids alone (0.43; 95%, 0.11-0.75; P=0.01).
• Mean (SD) infiltrate and/or scar size at 6 months was 3.03 mm in the standard therapy group, 2.85 mm in the early corticosteroid group, and 3.50 mm in the CXL group.
• Microbiological cure, defined as a negative culture at day 2 of treatment, was achieved in 88% of participants in the standard-therapy group, 85% of patients in the early-steroid group, and 84% of patients in the CXL group.
• Adverse events were rare, with 7% of participants experiencing corneal perforation or requiring therapeutic penetrating keratoplasty, 1% of participants had endophthalmitis, and 1 patient had an intraocular pressure elevation greater than 30 mmHg for 1 week.

The most common organisms isolated included Streptococcus pneumoniae and Pseudomonas aeruginosa, which were evenly distributed between treatment groups.

Limitations

• Most enrollees were non-contact lens wearing individuals in South India.
• Fungal, acanthamoeba, or herpetic ulcers were not included.
• Physicians were permitted to alter the antibiotic therapy as necessary following culture results or due to lack of response to therapy and the rate of antibiotic change was not reported between groups
• Only a small number of each bacterial organism was represented in this study, preventing conclusions regarding the best treatment for each organism
• As few participants had contact lens-related infections, different risk factors for infection may modify the interaction between infectious organisms and host responses, affecting efficacy of treatment

Conclusions

Neither topical corticosteroids nor CXL used as adjunctive treatments for bacterial corneal ulcers improved vision at 6 months compared with placebo, and adjunctive CXL was associated with increased scar size and more pain. There was also no significant difference in adverse events, such as corneal perforation, with use of topical corticosteroids or CXL. This suggests that these adjunctive therapies for the treatment of bacterial keratitis may not be superior to topical moxifloxacin monotherapy.

Clinical Pearls

• Topical corticosteroids do not have significant effect on clinical outcomes at 6 months in the treatment of culture-positive bacterial corneal ulcers, but they remain a safe adjunctive treatment option with no worse outcomes proven.
• Results in this South Indian population with predominantly S pneumoniae and P aeruginosa ulcers may not generalize well.
• Small numbers of perforation and therapeutic penetrating keratoplasty limited the power to detect a true effect, therefore CXL preventing melting and perforation cannot be ruled out.

Omega-3 Fatty Acid Supplementation for the Treatment of Dry Eye Disease, by the Dry Eye Assessment and Management (DREAM) Study Research Group

Objective

Dry eye disease (DED) is a widespread condition, affecting approximately 14% of adults in the United States. Previous clinical trials had demonstrated the efficacy of supplementation with poly-unsaturated fats, particularly omega-3 fatty acids, in mitigating inflammation. The DREAM trial sought to determine the effects of omega-3 fatty acid supplementation on dry eye disease.

Design

Multi-center, double-blind randomized clinical trial. Enrolled 535 patients with DED across 27 sites in the US. Subjects randomized to 2:1 ratio to receive active or placebo supplements for 12 months. 349 patients received active supplements, and 186 received placebo.

• Active Supplement Regimen: Active capsules contained 400 mg of eicosapentaenoic acid (EPA) and 200 mg docosahexaenoic acid (DHA), for a total daily dose of 2000 mg EPA and 1000 mg of DHA.
• Placebo Supplement Regimen: Placebo capsules contained 1000 mg of olive oil (68% oleic acid, 13% palmitic acid, 11% linoleic acid).

The regimen was reduced or suspended when patients reported gastrointestinal symptoms or when a contraindication to treatment with the full dose of active supplements developed. If these symptoms resolved, the patient could restart or increase the regimen.

• Visits were conducted at 3, 6, and 12 months after initiation of study. Participants were also contacted via telephone at 9 months to report adverse events.

Main outcome measure

1. Mean change from baseline in the Ocular Surface Disease Index (OSDI) score

Results

• OSDI scores decreased between baseline and 12 months in the active supplement group and in the placebo group (P<0.001), most of the decrease occurring in the first 3 months.
• The mean change in scores was -13.9+15.6 points in active supplement group and -12.5+18.2 points in placebo group, resulting in mean difference in change of -1.9 that was not statistically significant (95% CI, -5.0 to 1.1, p=0.21)

Limitations

According to the study design, participants were able to continue with or change their dry eye therapies during the study. Additionally, some have questioned the use of olive oil as a placebo, due to its potential anti-inflammatory properties.

Clinical Pearls

In patients with moderate-to-severe dry eye disease despite the use of other treatments, 12 months of supplementation with 3000 mg of omega-3 fatty acids did not contribute to a significant difference in improvement of dry eye disease symptoms and signs when compared to placebo.

Cornea Donor Study (CDS)

Background

Purpose: To evaluate whether survival with donors >65 years is comparable to younger donors in penetrating keratoplasty (PK) for moderate-risk endothelial disease.

Design: Prospective, multicenter, double-masked, randomized non-inferiority trial

Methods:

• Inclusion criteria: 1,090 patients age 40–80 years undergoing penetrating keratoplasty (PK) for Fuchs endothelial corneal dystrophy (FECD) or pseudophakic/aphakic corneal edema (PACE).
• Exclusion criteria: High-risk eyes (prior failed graft, chemical burns, herpes keratitis, uncontrolled glaucoma, uveitis) and low-risk eyes (e.g., keratoconus, stromal dystrophies without edema)
• Randomization: Randomization balanced assignment by surgeon and donor age (12–65 vs 66–75 years). Investigators and recipients were masked
• Primary outcome: Graft failure, defined as regraft or central opacity compromising vision ≥3 months. Rejection was classified as definite (endothelial rejection line) or probable (inflammatory signs without a line).

Key Findings

Donor Age and Graft Survival

• Cumulative probability of graft survival at 5 years was identical, and non-inferiority of donor age 12-65 compared to 66-75 years old was demonstrated. Difference in survival rate at 10 years was insignificant.

At 5 years, the cumulative probability of graft survival was 86% overall, identical between corneas from younger and older donors . The upper limit of the one-sided 95% CI was 4%, below the pre-specified non-inferiority margin of 8%, and adjustment for baseline ECD did not change the results. Donor age analyzed as a continuous variable showed no significant effect, although exploratory analyses suggested higher survival for the youngest donors (93% survival for donor age 12–40 years vs 85% survival for ages 41–75). At 10 years, survival was 77% for donors aged 12–65 versus 71% for donors aged 66–75. Survival remained stable for donors aged 34–71 (75%), with the greatest survival rate in donors aged 12–33 (96%) and the lowest in those aged 72–75 (62%), the latter decline emerging after year 6.

Endothelial Cell Loss and Morphometry

• Endothelial cell density (ECD) was significantly greater at 5 years in grafts from younger (12-65 years) compared to older (66-75 years) donors with a decline in ECD from baseline by about 70% across all donor ages.

The SMAS analyzed 347 eyes with clear grafts at 5 years. Median ECD declined by about 70% from baseline across all donor ages. Younger donors (12–65 years) had higher median 5-year ECD than older donors (824 vs 654 cells/mm²), corresponding to median losses of 69% and 75% (adjusted P = 0.04). At 10 years, substantial cell loss persisted at median ECD of 628 cells/mm² for younger donors and 550 cells/mm² for older donors . Higher baseline ECD and larger donor tissue size were associated with better long-term ECD, though variability was wide, with 24% of grafts falling below 500 cells/mm² and only 14% above 1,000 at 10 years. Morphometric analysis revealed that cell shape and variability independently predicted endothelial failure, beyond absolute ECD counts . A related analysis confirmed that low ECD alone strongly correlated with graft failure risk .

Causes of Graft Failure, Recipient Risk Factors, and Donor Risk Factors

• A preoperative diagnosis of PACE and glaucoma history, especially history of both glaucoma surgery and IOP-lowering drops, were the strongest predictors of graft failure at 5 years.

By 5 years, 135 grafts (12%) failed, including 102 regrafts (76%) and 33 failures without regraft (24%) . Major causes were graft rejection (48), endothelial decompensation (46), and primary donor failure (3). Other non-rejection failures included infection (15), persistent epithelial defects (6), glaucoma (3), epithelial downgrowth (2), corneal edema (1), thinning (1), hypotony (1), wound dehiscence (1), and refractive indications (8). Failure cause distribution did not differ by donor age. Among 1,090 participants, preoperative diagnosis and glaucoma history were the strongest predictors of failure . Risk was nearly four times higher with PACE compared with FECD (27% vs 7%). Glaucoma history markedly increased risk, especially when prior surgery and IOP-lowering medication use were both present. In FECD, lens status had little effect, whereas in pseudophakic/aphakic corneal edema, anterior chamber IOLs carried higher failure than posterior chamber IOLs. Recipient age, sex, diabetes, smoking, and graft size were not significant predictors. Elevated postoperative IOP, vitrectomy at transplant, and graft size reached significance in univariate but not multivariate models, largely confounded by diagnosis. Non-white/Hispanic recipients had higher failure risk than white non-Hispanic recipients, although subgroup sizes were small. Donor risk factors were less influential. Variables such as tissue retrieval method, cause of death, and preservation times showed no consistent effect . Importantly, donor diabetes was not associated with graft failure or with greater long-term endothelial cell loss at 10 years, however analyses at 10 years may be confounded by early failure of grafts from donors with diabetes that were not included in the analysis .

Corneal Thickness

• Analysis of corneal thickness measurements found that increased corneal thickness after PK independently predicted poorer long-term outcomes . Thicker grafts at 1, 5, and 10 years correlated with higher failure rates, providing a practical adjunct to ECD monitoring in predicting prognosis.

ABO Compatibility

• ABO compatibility was not a significant determinant of graft survival overall . In low-risk cases, ABO incompatibility also did not meaningfully alter outcomes .

Graft Rejection

• A definite rejection event was associated with lower graft survival rates compared to grafts without a rejection event.

Longer-term analyses confirmed the prognostic importance of rejection. By 10 years, a definite rejection event was strongly associated with subsequent graft failure: survival was 88% without rejection compared to 77% with definite rejection . A history of glaucoma, particularly with prior filtering surgery, markedly increased rejection risk (10-year incidence 35% vs 14%). At 5 years, rejection occurred in roughly one-quarter of grafts and was more frequent in PACE than FECD, as well as in female compared with male recipients . Lens status also influenced outcomes, with higher rejection in phakic FECD eyes versus pseudophakic eyes.

Clinical Implications

• Corneal tissue from donors up to 75 years can be used safely for PK in FECD and PACE.
• Five-year graft survival was identical between younger and older donor groups, and even at 10 years there was no significant difference (77% vs 71%), though ECL was somewhat greater in older tissue.
• Preoperative risk stratification is key, and surgical planning may benefit more from risk stratification based on recipient characteristics rather than based on donor variables. PACE was at fourfold higher risk of graft failure compared to FECD. Glaucoma history, particularly history of filtering surgery and use of IOP-lowering medications, was strongly predictive of poor survival. In PACE, anterior chamber intraocular lenses conferred greater failure risk than posterior chamber intraocular lenses.
• Preoperative ECD was not predictive of graft survival, whereas the 6-month ECD strongly predicted later graft survival. Increased corneal thickness during follow-up independently predicted poorer outcomes, complementing, but not replacing ECD monitoring.

Cornea Preservation Time Study (CPTS)

Background

Purpose: To provide high-quality evidence on whether preservation time (PT) up to 14 days affects endothelial cell loss (ECL) and graft success 3 years post Descemet stripping automated endothelial keratoplasty (DSAEK).

Design: Prospective, multicenter, double-masked, randomized, noninferiority trial.

Methods:

• Inclusion criteria: Patients age 30 to 90 with Fuchs’ endothelial corneal dystrophy (FECD) or pseudophakic/aphakic corneal edema (PACE) [3].
• Exclusion criteria: Patients with prior failed PKP or DSAEK, tube shunts, uncontrolled glaucoma or uveitis, anterior chamber intraocular lenses, and peripheral anterior synechiae measuring more than one-fourth of the anterior chamber angle were excluded.
• Randomization: Patients were randomized 1:1 to receive donor corneas preserved in Optisol-GS for ≤7 days or 8–14 days prior to transplantation. Surgeons were required to have prior experience with at least 50 DSAEK cases, a primary donor failure rate of less than 3%, and they were masked to PT until surgery.
• Primary outcome: Graft success at 3 years defined as a clear graft without need for regrafting. Graft failure was defined as regraft for any reason, failure to clear 8 weeks postoperatively, or initially clear graft that became and remained cloudy for 90 days.

Key Findings

Graft Success and PT

• Success rate at 3 years after DSAEK were high regardless of PT, however noninferiority of longer PT was unable to be demonstrated.

At 3 years, the cumulative probability of graft success was not significantly different between groups (95.3% in the 0-7 day group; 92.1% in the 8-14 day group), however results did not meet criteria for noninferiority of the 8-14 day group [3]. When further divided into subgroups, 3-year graft success was lowest for PT of 12-14 days compared to other preservation time subgroups. The risk of graft failure was significantly higher in the 8–14 day group. Subgroup analysis further showed that the 3-year graft success was lowest for preservation times of 12–14 days compared to graft success for PT of ≤4 days, 5–7 days, and 8–11 days.

Endothelial Cell Loss

• Longer PT was associated with significantly greater endothelialcell loss (ECL), however, the effect of PT on endothelial cell loss was comparable from 4-13 days. Greatest ECL occurred within the first 6 postoperative months.

At 3 years post DSAEK, ECD declined by 37% in the 0–7 day group and by 40% in the 8–14, with eac additional day of PT corresponding to am ean decrease of 15 cells/mm². Findings at 4 years post DSEAK were similar.

Graft Rejection

• Overall graft failure rate from rejection was low, and younger recipient age was the only significant factor in risk of rejection.

The 3-year cumulative probability of a definite rejection episode was 3.6%, with younger recipient age being the only significant factor in risk of rejection. Half of definite rejection episodes occurred within the first postoperative year. The overall graft failure rate from rejection in CPTS was low at 1%. Donor age, PT, immunization within 3 months of follow-up, and graft size, were not significantly associated with rejection risk. Contrary to the Cornea Donor Study (CDS), which found that PACE, prior use of glaucoma medications, and glaucoma filtering surgery were associated with higher risk of graft rejection following PK, CPTS found that prior use of glaucoma medications and less complex glaucoma surgery (e.g. laser trabeculoplasty, trabeculectomy) were not associated with higher risk of graft rejection following DSAEK. However, notably CPTS did not include tube shunt surgeries. Grafts that had experienced a rejection episode but were ultimately clear at 3 years had significantly higher ECL than grafts that had no history of rejection (48% vs 38%, P=0.03) but did not substantially impact overall graft survival.

Donor, Recipient, and Operative Factors

• Donor diabetes, lower screening ECD, a recipient diagnosis of PACE, and operative complications were significantly associated with lower ECD and reduced graft survival at 3 years.

Impact on ECD Mean endothelial cell loss (ECL) was 47% in recipients of corneas from diabetic donors compared with 43% without diabetes. PACE was associated with greater ECL compared with FECD (53% vs 44%). Preoperative ECD was not associated with LEGF but lower ECD at 6 months was strongly predictive (p<0.001) Lower baseline intraocular pressure was initially associated with reduced 3-year ECD, but this was not significant after excluding eyes with prior glaucoma treatment, suggesting baseline IOP did not independently affect long-term endothelial survival when glaucoma was well controlled and anterior chamber anatomy was well preserved. Notably, intraoperative complications were strongly associated with worse outcomes (55% ECL loss with complications, 44% without complications.

Impact on graft survival The overall 3-year graft success rate in CPTS was 94.1% . Donor diabetes and operative complications were independent predictors of failure. At 3 years, graft success was 95.0% for non-diabetic donors versus 90.3% for diabetic donors, and 94.6% without operative complications versus 79.5% with complications. PACE was the principal recipient factor associated with late graft failure; 3-year success rates were 83.7% in PACE and 94.3% in FECD. Other donor, recipient, and operative characteristics including donor age, sex, storage solution, cause of death, lenticule thickness, insertion method, incision size, concomitant cataract surgery, and recipient age were ultimately not independent risk factors for endothelial survival. With regards to graft success, eyes with prior glaucoma surgery showed success rates (80% vs 94%), but this did not meet the significance threshold in CPTS, and tube shunt cases were excluded. In summary, donor diabetes, lower screening ECD, a recipient diagnosis of PACE, and operative complications were significantly associated with lower ECD at 3 years. Donor diabetes, operative complications, and PACE were likewise independent predictors of reduced graft survival. Operative complications were the most clearly modifiable factor, underscoring the importance of careful surgical technique.

Postoperative Graft Attachment and Intraocular Pressure (IOP)

• Donor diabetes, greater pre-lamellar dissection donor central corneal thickness, and intraoperative complications were associated with higher rates of graft dislocation, and graft dislocation was associated with higher risk of graft failure at 3 years postoperatively. Early acute IOP elevation was associated with higher risk of graft failure.

Eight percent of eyes had at least one graft dislocation (GD) postoperatively. Donor diabetes, greater pre-lamellar dissection donor central corneal thickness, and intraoperative complications (OR, 2.97) were associated with a higher risk of GD. After accounting for donor history of diabetes, PT, recipient diagnosis, operative complications and surgeon, GD was a major determinant of long-term outcomes for graft failure compared with eyes without GD at 3 years postoperatively. These results highlight that even partial detachment carried an elevated risk of failure compared to fully attached grafts, supporting a “dose–response” effect for the severity of dislocation. GD was also associated with significantly lower 3-year postoperative ECD compared with eyes without GD [10]. Graft dislocation, but not partial detachment with interface fluid, was associated with worse ECD in clear grafts at 3 years, which the authors suggested may adversely affect longer-term graft survival. The 3-year graft success rate for 24 eyes (75%) with acute IOP elevation was significantly lower than in eyes that did not experience the acute increase in IOP (94.1%).There was a 3.4-fold higher risk of graft failure (HR, 3.42), although it did not significantly affect the mean 3-year ECD. The authors noted that this apparent discrepancy was likely due to a selection effect: grafts severely damaged by acute IOP spikes tended to fail early, leaving only surviving clear grafts available for ECD analysis at the 3 year mark [11] By contrast, elevated IOP beyond 1 month postoperatively occurred in 23% of eyes with functioning grafts, but was not significantly associated with graft success or ECD at 3 years.

Infections and Donor Rim Cultures

• Longer PT was not associated with increased rim culture positivity for fungal or bacterial growth, and overall rate of infection was low.

Donor rim cultures were performed in 59% of eyes. Positive fungal growth occurred in 2.5% of 0–7 day PT corneas and 1.3% of 8–14 day PT corneas, while positive bacterial growth occurred in 1.5% and 1.0%, respectively, with insignificant differences in percentage of positive cultures between PT groups. Surgeon-prepared tissue carried a higher risk of positive fungal cultures compared with eye bank–prepared tissue, while younger donor age and accidental death were associated with positive bacterial cultures. Postoperative infection was rare: fungal keratitis developed in 1 of 15 recipients of a cornea with a positive fungal culture (6.7%), and no infections occurred following positive bacterial cultures. With two additional infections in eyes without rim culture performed, the overall incidence across the entire CPTS cohort was 0.15% for fungal and 0.08% for bacterial infection. Longer PT was not associated with increased rim culture positivity or infection risk, and the overall rate of infection was very low.

Surgeon Attitudes and Policy Impact

• The proportion of surgeons willing to accept donor corneas preserved for >7 days and eye bank-reported mean PT increased after publication of CPTS.

A follow-up survey of American Academy of Ophthalmology members with a cornea interest assessed the impact of CPTS on surgeon attitudes toward PT. In 2012, 364 of 1,609 surgeons (22.6%) responded, and in 2018, 297 of 1,872 surgeons (15.9%) responded. The proportion of surgeons willing to accept donor corneas preserved for more than 7 days significantly increased from 32% in 2012 to 46% in 2018. A significant change was observed among surgeons with more than 10 years of experience, with acceptance rising from 28% to 47% over this period. Consistent with these shifts, eye bank-reported mean PT increased from 4.6 days in 2010 to 5.1 days in 2018, and the proportion of endothelial keratoplasty (EK)-intended donor corneas preserved for more than 7 days rose from 3% to 9%. These results suggest that CPTS influenced both surgeon acceptance patterns and eye-bank distribution practices.

Clinical Implications and Summary

• Donor corneas can be safely preserved for up to 11 days with little influence on graft success and comparable ECL from 4–13 days
• Overall infection incidence was low
• Donor diabetes was consistently associated with lower 3-year ECD and higher graft failure risk, while PACE (vs FECD) conferred greater ECL and an elevated risk of late failure. Other variables including donor age, graft diameter, and insertion method were not independent predictors of outcomes within CPTS criteria.
• Surgical technique and early postoperative events emerged as the most clearly modifiable factors. Operative complications were strongly linked to both reduced ECD and increased failure risk. Graft dislocation (GD), seen in 8% of eyes, carried a nearly eight-fold higher failure risk, with a clear dose–response relationship by detachment severity.
• 6-month postoperative ECD, not preoperative ECD, was a critical predictor of late endothelial graft failure at 5 years, reinforcing the need to protect the endothelium early after surgery

Diabetes Endothelial Keratoplasty Study

Objectives

To investigate whether 1-year graft success rate in Descemet Membrane Endothelial Keratoplasty (DMEK) is similar between grafts from donors with versus without diabetes.

Design

Randomized clinical trial between February 2022 to July 2024 including 1097 participants (324 bilateral) age 30 to 91 years old with diagnosis of Fuchs endothelial corneal dystrophy (FECD), pseudophakic corneal edema, or previously failed DSAEK or DMEK across 28 clinical sites.

Diabetes severity was graded on a scale of 1-5 by a sum of points given as follows: 1 for any history of diabetes, 1 for BMI > 30, 1 for hypertension, and automatically 2 points given for any of diabetes history over 10 years, insulin dependence outpatient, end organ damage (including conditions such as peripheral neuropathy, peripheral vascular disease, renal failure, stroke, myocardial infarction, retinopathy). Skin samples were taken at autopsy for analysis of advanced glycation end products.

Patients with failed penetrating keratoplasty, uncontrolled uveitis, prior tube shunt surgery, uncontrolled glaucoma, aphakic corneal edema, anterior chamber IOL, among other characteristics were excluded.

Main outcome measures

The main outcome measure was graft failure at 1 year defined as occurrence of one of the following: cornea that required regrafting for any reason; cornea that remained cloudy without clearing either if 1) cornea was cloudy at POW1 that did not clear by 7 weeks OR 2) cornea that was initially clear postoperatively but became and remained cloudy for at least 3 months.

Results

No significant difference was found in 1-year graft success rate between grafts from donors with versus without diabetes. Success rate was high (>96%) regardless of donor diabetes status. When success rate from donors from each diabetes severity category was compared against success rate from donors without diabetes, no significant difference was found. Endothelial cell loss (ECL) and endothelial cell morphometry were not affected by donor diabetes status.

Limitations

One year follow-up period, mostly FECD patients.

Conclusions

Tissue across the full spectrum of donor diabetes severity is usable. This has implications for expanding the donor tissue pool especially as prevalence of diabetes increases.

References

Isa.Mohammed