Highlight
The Diabetes Endothelial Keratoplasty Study (DEKS) identified preoperative donor endothelial cell density (ECD), donor endothelial cell damage assessed via trypan blue staining, the recipient’s diagnosis, and intraoperative complications as significant predictors of endothelial cell density at one year following Descemet membrane endothelial keratoplasty (DMEK).
Higher endothelial cell loss correlated with increased donor endothelial cell damage, non-Fuchs endothelial corneal dystrophy (FECD) recipient diagnoses, and operative complications. These findings emphasize the importance of meticulous donor tissue preparation and operative technique to optimize long-term graft endothelial survival.
Study Background
Descemet membrane endothelial keratoplasty (DMEK) is a preferred surgical technique for treating corneal endothelial dysfunction, notably in patients with Fuchs endothelial corneal dystrophy (FECD). Maintaining adequate endothelial cell density (ECD) post-transplant is critical for graft clarity and function, as these cells do not regenerate extensively.
Despite DMEK’s advantages over prior endothelial keratoplasty procedures, endothelial cell loss remains a concern, potentially compromising graft survival and visual outcomes. Prior studies have indicated that donor characteristics, recipient pathology, and surgical factors might influence postoperative ECD, but comprehensive multicenter, prospective data have been limited.
The Diabetes Endothelial Keratoplasty Study (DEKS), a randomized, multicenter trial, provided an opportunity to investigate factors influencing ECD at one year after DMEK in a large cohort. Understanding these factors can improve donor tissue selection, surgical preparation, and recipient management to enhance graft longevity.
Study Design
The DEKS prospective cohort study was a predefined secondary analysis from a multicenter, double-masked, randomized clinical trial. It included 1,421 eyes undergoing DMEK, of which 1,274 eyes were analyzed at one year postoperatively with evaluable central endothelial images.
Donor corneas originated from donors with or without diabetes, and endothelial images were obtained preoperatively, assessing ECD centrally. The main outcome measured was ECD at one year after DMEK. The study adjusted for surgeon variability using mixed-effect models and accounted for correlations between fellow eyes.
Key variables analyzed included donor preoperative ECD, extent of endothelial cell damage reported by eye banks following lenticule preparation (quantified by trypan blue staining), recipient diagnosis (FECD versus other indications), and presence of operative complications.
Key Findings
Donor Factors: Lower donor preoperative ECD was significantly associated with reduced ECD at one year; every decrease of 500 cells/mm2 preoperatively predicted a loss of approximately 392 cells/mm2 at one year (99% CI -447 to -336). Notably, endothelial cell damage assessed by trypan blue staining after lenticule preparation independently predicted lower postoperative ECD. Eyes with 10–20% endothelial cell damage had a mean reduction of 161 cells/mm2 compared to those with less than 5% damage.
Operative Factors: Intraoperative complications were associated with a substantial decrease in ECD, with an estimated effect of -308 cells/mm2 (99% CI -450 to -166) at one year. Eyes with complications lost an average of 42% of their endothelial cells compared to 27% in uncomplicated cases.
Recipient Factors: Recipient eyes with diagnoses other than FECD exhibited greater ECD loss, with a decrease of 191 cells/mm2 at one year compared to FECD recipients. Endothelial cell loss was 40% in non-FECD indications versus 28% in FECD cases.
Overall Endothelial Cell Loss: When stratified by preoperative donor ECD, one-year cell loss ranged from 26% to 30%. Greater cell loss was observed with increasing levels of endothelial cell damage and in the presence of intraoperative complications or alternative recipient indications.
Expert Commentary
This study provides valuable evidence that meticulous assessment of donor endothelial quality beyond cell counts—specifically, quantifying endothelial cell damage after tissue preparation—can serve as an independent marker to predict postoperative graft survival. The use of trypan blue staining adds an important dimension to donor screening protocols and may guide eye banks and surgeons in selecting optimal tissues, particularly for vulnerable recipient populations.
The association of greater endothelial loss in non-FECD cases may reflect underlying host factors, including more complex disease processes or differences in recipient ocular microenvironments that influence graft survival. Intraoperative complications remain a significant modifiable risk factor, underscoring the need for surgical expertise, careful tissue handling, and operative protocols to reduce trauma to the delicate endothelial layer.
Limitations include the predominance of FECD cases and white participants, which may affect generalizability. Future studies should also evaluate longer-term outcomes beyond one year and explore interventions to mitigate endothelial damage during preparation and surgery.
Conclusion
The Diabetes Endothelial Keratoplasty Study identifies donor endothelial cell damage as a novel, independent predictor of endothelial cell density at one year post-DMEK, alongside donor ECD, recipient diagnosis, and intraoperative complications. These results highlight the critical need for optimized tissue assessment and surgical precision to enhance graft survival and patient outcomes after DMEK.
Eye banks and surgeons should incorporate quantitative endothelial viability assessments such as trypan blue staining into donor screening and preparation protocols. Further, recognizing higher cell loss risks in non-FECD recipients may inform tailored postoperative management strategies. Overall, this study advances our understanding of factors influencing transplant success in corneal endothelial keratoplasty.
Funding and Clinical Trial Registration
The Diabetes Endothelial Keratoplasty Study Group conducted and funded this multicenter, randomized trial. Detailed funding sources and trial registry identifiers were not specified in the analyzed publication.
References
1. Patel SV, Benetz BA, Price MO, et al. Factors Associated with Endothelial Cell Density at One Year in the Diabetes Endothelial Keratoplasty Study. Ophthalmology. 2026 Jul 30; PMID: 42532315.
2. Price MO, Price FW Jr. Descemet’s membrane endothelial keratoplasty: early challenges and future directions. Curr Opin Ophthalmol. 2013;24(4):329-35.
3. Melles GRJ, Ong TS, Ververs B, van der Wees J. Descemet membrane endothelial keratoplasty (DMEK). Cornea. 2006;25(8):987-90.
