Endothelium-in Versus Endothelium-out Insertion With Descemet Membrane Endothelial Keratoplasty

To determine whether graft orientation during insertion affects Descemet membrane endothelial keratoplasty (DMEK) outcomes.

This was a retrospective analysis of 754 consecutive DMEK cases performed by 2 experienced surgeons to treat Fuchs dystrophy. Both surgeons used an intraocular lens insertor. One surgeon always inserted the tissue scrolled endothelium outward (group 1, n = 245). The other surgeon tested 3 methods: endothelium-outward scroll configuration (group 2, n = 161), endothelium-inward trifold configuration (group 3, n = 172), and trifold configuration with concurrent use of an anterior chamber maintainer (group 4, n = 176). The main outcome measures were rebubbling rate, regrafting or failure within 6 months, and 6-month endothelial cell loss. The tissue unfolding time from graft insertion to air fill was measured in a subset of 120 cases by 1 surgeon.

The rebubbling rates were comparable across groups (ie, 12%, 10%, 10%, and 13% for groups 1, 2, 3, and 4, respectively, P = 0.21). The 6-month graft failure/replacement rates were comparable across groups (ie, 0.8%, 1.2%, 2.3%, and 0.6%, respectively, P = 0.18). Similarly, the 6-month endothelial cell loss did not differ significantly between groups (ie, 28% ± 11%, 30% ± 13%, 28% ± 15%, and 27% ± 13%, respectively, P = 0.019). In the subset analysis, the tissue unfolding time was similar for scroll and trifold configurations (6.0 ± 3.5 vs. 5.4 ± 3.0 minutes, respectively, P = 0.43).

The outcomes were similar for endothelium-out and endothelium-in (trifold) insertion methods with DMEK, suggesting that the choice is a matter of surgeon preference.