Abstract
PURPOSE: To identify pre/perioperative variables that shape endothelial cell loss (ECL) after uncomplicated Descemet membrane endothelial keratoplasty (DMEK). METHODS: This retrospective study included all consecutive patients with Fuchs endothelial corneal dystrophy who underwent DMEK surgery without perioperative or postoperative complications in 2015-2023 and were followed for 12 months. To identify covariates that predicted 12-month ECL, primary hierarchical multivariable analysis was conducted with 12 variables: patient age and sex; donor age; preoperative axial length, visual acuity, central corneal thickness, and graft endothelial cell density; endotamponade with sulfur hexafluoride (SF6) or air; triple-DMEK or pseudophakic-DMEK; operative time; graft marking; and rebubbling. RESULTS: 137 eyes (112 patients) were included. Multivariable analysis showed that SF6 predicted 13.6 ± 3.4% greater ECL vs. air (p < 0.0001) and accounted for 10% of total ECL variation. Longer operative time and multiple (≥2) rebubbling also predicted 0.4 ± 0.7% (p = 0.046) and 11.7 ± 5.1% (p = 0.02) higher ECL, respectively. SF6 significantly reduced rebubbling on univariable analysis (13% vs. 41% for air, p = 0.01). CONCLUSIONS: SF6 use for endotamponade may increase ECL after DMEK. There is an urgent need for randomized controlled trials that estimate the relative disadvantages (endothelial toxicity) and advantages (less bothersome rebubbling) of SF6. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT02535819.