We thank Galvis et al for their interest in our article. We are happy to clarify certain points as requested. In our study, patients' ages ranged from 39 to 84 years (mean 63). As reported, this was not found to correlate with time to clearance. Lens status was not an inclusion or exclusion criterion. Cases 2 and 6 were pseudophakic on enrollment; the remainder of the cases were phakic. Cataract surgery was not performed in any patient in combination with descemetorhexis or during the follow-up period.
We note the authors' concern regarding endothelial cell counts and potential long-term stability. We would point out that the cell counts quoted by Galvis et al are central cell counts. Before performing descemetorhexis in all cases, apart from case 11, central cell counts were unrecordable because of the presence of dense guttae. We would argue that these cell counts should therefore be regarded in a different manner than we are used to as transplant surgeons. These numbers represent an increase in endothelial density from a baseline of essentially zero. A more important question we believe is whether this transition of cells to the center has exhausted the peripheral reserve. A cell count at 12 months of 500/mm2 with a peripheral density of over 2000/mm2 may be quite stable and expected to increase further (our pilot case of descemetorhexis for Fuchs dystrophy has a stable endothelial cell mosaic and a clear cornea at 3.5 years). A cell count of 500/mm2 over the entire cornea may be of more concern. It is our clinical impression that in reporting cell counts after this procedure, the peripheral density, in particular, the superior zone, may be the more important indicator of long-term stability, and we intend to include this in all future publications of similar cohorts.
The authors' enthusiasm for ripasudil as a potential endothelial panacea is understandable. The agent holds promise as an accelerant of endothelial healing; we believe that it is yet to be determined whether it can prolong corneal clarity in the presence of low cell counts. Studies demonstrating this will of course take far longer. Although their downstream effects are many, the Rho-kinase inhibitors act primarily on the internal cellular cytoskeleton to effect morphologic changes in cells.1–3 It is through this action on the trabecular meshwork, for example, that an intraocular pressure–lowering effect is felt to be achieved. Our own in vitro analysis demonstrates an increase in markers of mesenchymal transformation such as vimentin (Fig. 1). In addition, we have observed endothelial cell morphologic changes in vitro and in vivo, as did the Kyoto group.1 It is our early impression that ripasudil acts to transform endothelial cells into a more motile phenotype, more able to overcome local inhibitory factors to migrate across denuded areas.4 In this sense, it may prove to be the endothelial equivalent of autologous serum or tarsorrhaphy that we so commonly use to achieve epithelial closure.
Although the drug has passed phase III clinical trials and is to market in 1 country, we would approach its use for this new indication with some caution. Use in a prospective, controlled clinical trial setting with detailed patient consent, objective outcome criteria, and longer follow-up would still seem the most appropriate. We are attempting to encourage mitosis and mesenchymal transformation of the endothelial monolayer. In vitro, this transformation can be difficult to reverse, which would be highly undesirable in vivo. Many systemic uses of ROCK inhibitors have also been proposed, with possible off-target effects.