We thank Price and Price for their comments on our article on corneal clearance despite graft detachment after Descemet membrane endothelial keratoplasty (DMEK). Since the submission of our article, there have been several other cases with a similar postoperative course, that is, the endothelialization of recipient posterior stroma in the presence of a detached graft ( Figure ). Although the endothelial cell density is closely monitored in all cases, the postoperative intervals may yet be too short to anticipate on any further increase in cell density, that is, cell counts exceeding 300 to 500 cells/mm 2 .
As discussed in our article and suggested by Price and Price, the positioning of the graft during surgery may have been upside down, with the endothelium facing the recipient posterior stroma. The outward curl of the donor Descemet membrane into the anterior chamber certainly would be indicative for a reversed orientation. However, the position of the graft is checked meticulously during surgery, and in most cases, the surgical videos seemed to confirm that during surgery, the Descemet graft was positioned right side up, that is, with inward curls toward the stroma. If so, this would leave 2 possibilities:
First, the tendency of an in situ donor Descemet membrane to curl inward or outward may not reflect that of a so-called free graft in an in vitro situation or within the anterior chamber. Particularly when the graft is partially attached, the free floating part of the tissue may be forced to curl in the opposite direction, like the bimorphic position with the lid of a can: it is either concave or convex. Second, it can not be ruled out that some Descemet grafts may roll up with the endothelium turned inward. Although the curl direction during the formation of a Descemet roll during preparation of the graft (i.e., stripping off the Descemet membrane of a corneoscleral rim) is monitored closely in our eyebank, the clinical cases presented may suggest that the roll direction may flip, for example, because of irrigation during surgery.
In the literature, there may be conflicting reports on endothelial wound healing, regeneration, or both. In our laboratory, we found that endothelial wound healing capacity in experimental settings may be different from that in vivo. Nevertheless, we clinically observed that the mere implantation of a free-floating donor Descemet membrane with its endothelium (tentatively referred to as Free DMEK) may be an alternative and less invasive treatment option for endothelial disorders. Although Free DMEK may provide 20/20 visual acuity within 6 to 12 months after surgery, we currently are trying to further identify the underlying mechanism to standardize the procedure better, before commencing a (multicenter) clinical trial.