DMEK Outcomes Using Tissue from Donors with Diabetes

Results from the Diabetes Endothelial Keratoplasty Study (DEKS) show that Descemet membrane endothelial keratoplasty (DMEK) in eyes with uncomplicated corneal endothelial dysfunction has a very high success rate at one year, irrespective of donor diabetes status. In addition, analyses showed that neither endothelial cell loss (ECL) nor morphometric changes were related to donor diabetes status or severity.

The findings were published online in JAMA Ophthalmology.^1,2 Jonathan H Lass MD, the DEKS study chair, discussed the results at a recent conference.

“DMEK is the most common keratoplasty procedure performed in the United States, and the findings from the DEKS support the use of tissue across the full spectrum of donor diabetes severity,” Dr Lass said. “Acceptance of these results by the cornea community should expand the pool of donor tissue.”

A marked rise in the proportion of corneas from diabetics in the donor pool, along with studies reporting poorer outcomes using this tissue, provided the impetus for the DEKS. Sponsored by the National Eye Institute (NEI), the DEKS is a randomised trial that enrolled patients aged 30 to 90 years undergoing DMEK for uncomplicated Fuchs’ endothelial corneal dystrophy (FECD), pseudophakic cornea oedema, or failed endothelial keratopathy. It involved 46 surgeons from 28 clinical sites and donor corneas provided by 13 US eye banks. Two-thirds of the grafts were from normal donors and one-third from donors with diabetes as confirmed by postmortem HbA1c testing.

The study group included 1,421 eyes of 1,097 patients, of which 96% completed the one-year follow-up visit. Comparisons between characteristics of the diabetic and non-diabetic donor corneas and the non-diabetic and diabetic recipient groups showed they were well matched.

The primary endpoint analysis showed that at one year, the graft success rate was 96.3% in eyes receiving a non-diabetic donor and 97.1% in the diabetic donor group. Analyses of ECL were performed in a sub-cohort of 1,274 eyes with clear grafts at one year and analysable images at one month and/or one year. Baseline donor and recipient characteristics were similar in the diabetic and non-diabetic groups. In both the recipient groups receiving donors without and with diabetes, the ECL rate was 24% at one month and 28% at one year.

“The ECL rate was also 28% in both the ‘double whammy’ subgroup consisting of cases with a donor and recipient with diabetes compared to the subgroup with both recipients and donors without diabetes,” Dr Lass said.

ECL was also not affected by diabetes severity, although consistent with previous studies, it was significantly higher (40%) in eyes with pseudophakic corneal oedema and failed endothelial keratoplasty compared to 28% with eyes with FECD.

Additional reports based on analyses at one year are forthcoming.

“Along with donor diabetes status, the DEKS prospectively collected over 60 preoperative and postoperative factors that will be examined in secondary analyses to identify predictors of DMEK outcomes,” Dr Lass said.

He also announced that the NEI has provided funding to support five-year follow-up of DEKS participants to validate the long-term findings regarding donor diabetes status and DMEK outcomes.

“And as a novel aspect, a genetics component has been added to assess if donor and recipient FECD and diabetes genetics affect longer term endothelial cell loss and graft failure,” Dr Lass said.

Dr Lass spoke at AAO 2025 in Orlando, US.

Jonathan H Lass MD is Study Chair of the DEKS and Charles I Thomas Professor, Department of Ophthalmology and Visual Sciences, Case Western Reserve University and University Hospitals Eye Institute, Cleveland, Ohio, US. jonathan.lass@uhhospitals.org

1. Price FW Jr, et al. JAMA Ophthalmol, 2025 October 17; 143(12): 1043–1051. Online ahead of print.

2. Lass JH, et al. JAMA Ophthalmol, 2025 Oct 17; 143(12): 1053–1060. Online ahead of print.