"Endothelium-Out" and "Endothelium-In" Descemet Membrane Endothelial Keratoplasty (DMEK) Graft Insertion Techniques: A Systematic Review With Meta-Analysis

Background: We evaluated the visual outcomes and complications of "endothelium-out " and "endothelium-in " Descemet membrane endothelial keratoplasty (DMEK) graft insertion techniques. Materials and Methods: Electronic searches were conducted in CENTRAL, Cochrane databases, PubMed, EMBASE, . Study designs included clinical trials, comparative observational studies, and large case series >= 25 eyes). PRISMA guidelines were used for abstracting data and synthesis. Random-effects models were employed for meta-analyses. Results: 21,323 eyes (95 studies) were included. Eighty-six studies reported on "endothelium-out " techniques; eight studies reported on "endothelium-in " techniques. One study compared "endothelium-out " to "endothelium-in " techniques. Eighteen "endothelium-out " studies reported that 42.5-85% of eyes achieved best-corrected visual acuity (BCVA) >= 20/25 at 6 months; pooled proportion of eyes achieving BCVA & GE;20/25 at 6 months was 58.7% (95% CI 49.4-67.7%,15 studies). Three "endothelium-in " studies reported that 44.7-87.5% of eyes achieved BCVA of >= 20/25 at 6 months; pooled proportion of eyes achieving BCVA >= 20/25 at 6 months was 62.4% (95% CI 33.9-86.9%). Pooled mean endothelial cell loss was lower in the "endothelium-in " studies (28.1 +/- 1.3%, 7 studies) compared to "endothelium-out " studies (36.3 +/- 6.9%,10 studies) at 6 months (p = 0.018). Graft re-bubbling rates were higher in the "endothelium-out " studies (26.2%, 95% CI 21.9-30.9%, 74 studies) compared to "endothelium-in " studies (16.5%, 95% CI 8.5-26.4%, 6 studies), although statistical significance was not reached (p = 0.440). Primary graft failure rates were comparable between the two groups (p = 0.552). Quality of evidence was considered low and significant heterogeneity existed amongst the studies. Conclusion: Reported rates of endothelial cell loss were lower in "endothelium-in " DMEK studies at 6 months compared to "endothelium-out " studies. Outcomes of "endothelium-in " techniques were otherwise comparable to those reported in "endothelium-out " studies. Given the technical challenges encountered in "endothelium-out " procedures, surgeons may consider "endothelium-in " techniques designed for easier intra-operative DMEK graft unfolding. "Endothelium-in " studies evaluating outcomes at longer time points are required before conclusive comparisons between the two techniques can be drawn.