24 Effect of graft preparation techniques on clinical outcomes after Descemet membrane endothelial keratoplasty (DMEK).

Purpose: To compare the clinical outcomes after Descemet membrane endothelial keratoplasty (DMEK) for grafts prepared by the manual no-touch peeling technique and grafts prepared by a modified liquid bubble technique.

Material And Methods: For this study, 236 DMEK grafts were included that were prepared at Amnitrans EyeBank Rotterdam by experienced eye bank personnel. 132 grafts were prepared by using the 'no-touch' DMEK preparation technique and 104 grafts by using a modified liquid bubble technique. The liquid bubble technique was modified to render it a no-touch technique while maintaining the ability to save the anterior donor button as a potential Deep Anterior lamellar keratoplasty (DALK) or Bowman layer (BL) graft. DMEK surgeries were performed at Melles Cornea Clinic Rotterdam by experienced DMEK surgeons. All patients underwent DMEK for Fuchs endothelial dystrophy. Average patient age was 68 (±10) years and average donor age was 69 (±9) years with no difference between the two groups. Endothelial cell density (ECD) was evaluated after graft preparation by light microscopy in the eye bank and at 6-month postoperatively by specular microscopy.

Results: Endothelial cell density (ECD) decreased from 2705 (±146) cells/mm2 (n=132) before to 1570 (±490) cells/mm2 (n=130) at 6 months postoperatively for grafts prepared by the no-touch technique. For grafts prepared by the modifiedliquid bubble technique, ECD decreased from 2627 (±181) cells/mm2 (n=104) before to 1553 (±513) cells/mm2 (n=103) after surgery. Postoperative ECD did not differ for grafts prepared by the two techniques (P=0.79). Central corneal thickness (CCT) decreased from 660 (±124) µm to 513 (±36) µm postoperatively in the no-touch group and from 684 (±116) µm to 515 (±35) µm postoperatively in the modified liquid bubble group, with no postoperative CCT difference between groups (P=0.59). In total 3 eyes underwent re-surgery within the study period (n=2 (1.5%) in the no-touch group, n=1 (1.0%) in the liquid bubble group; P=0.71) and 26 eyes required a re-bubbling procedure for incomplete graft adherence (n=16 (12%) in the no-touch group, n=10 (10%) in the liquid bubble group; P=0.37).

Conclusion: Clinical outcomes after DMEK are comparable for grafts prepared by either the manual no-touch peeling technique or the modified liquid bubble technique. While both techniques are safe and useful techniques to prepare DMEK grafts, the modified liquid bubble technique offers advantages for corneas with scars.