Hydrops after corneal perforation during manual deep anterior lamellar keratoplasty for keratoconus.

Abstract

In our recently published article ‘Are Descemet membrane ruptures the root cause of corneal hydrops in keratoconic eyes?’ (Parker et al. 2019), we described how hydrops formation in keratoconus (KC) may be subjected to a more complex mechanism than currently hypothesized, that is, a hydrops would result from a break in Descemet membrane (DM) in the background of abnormal collagen of the corneal stroma in KC (Dua et al. 2015; White et al. 2017; Loh et al. 2018; Parker et al. 2019). In eyes with Fuchs endothelial corneal dystrophy and coexisting KC, removal of DM during descemetorhexis never resulted in any sign of hydrops formation, while a stromal perforation into the anterior chamber during Bowman layer transplantation consistently produced a significant hydrops (Parker et al. 2019). As such, a rupture of the most posterior stromal layers, and not so much a defect at the level of DM alone seems to be key for corneal hydrops to occur. For the implantation of an isolated Bowman layer graft, we have been using a manual dissection technique to create a stromal pocket to accommodate the transplant (Parker et al. 2019). This method of dissection is also routinely used for manual deep anterior lamellar keratoplasty (DALK) (Melles et al. 1999). Therefore, it would stand to reason that a perforation during DALK in KC eyes would also result in hydrops formation. Hence, we retrospectively screened the records of a series of 25 eyes of 25 patients [mean age 33 (±14) years; range 16–61 years; 15 male and 10 female] that underwent manual DALK for KC for corneal perforation and subsequent hydrops formation. Five eyes of five patients with advanced KC [mean age 30 (±5) years; range 23–37 years; four male and one female], all phakic, had experienced a central corneal perforation during surgery (Table 1). In four eyes, the DALK procedure could be completed with the positioning of an anterior donor corneal button, while the surgery was aborted in one eye. The latter eye clearly showed a similar picture as that seen in eyes with a central perforation during Bowman layer transplantation (Parker et al. 2019). Interestingly, the other four eyes, in which a donor corneal button was sutured in place, showed swelling, displacement and/or disruption of the most posterior corneal stroma left in situ, while the anatomy of the overlying graft – although swollen – appeared normal. Our observation may further support the hypothesis that, apart from a break in the posterior cornea, an abnormal host mid-stroma may be important in hydrops formation (Dua et al. 2015), because host tissue showed significant swelling while the donor corneal button did not. Furthermore, since peripheral intraoperative perforations usually seem unaccompanied by significant tissue distortion during and after surgery, the abnormal stromal anatomy may largely be limited to the central cornea and/or the actual ‘cone’. In conclusion, these observations support our recent findings that a combined defect in DM and the anatomically abnormal posterior stroma, rather than solely a break in DM, is associated with corneal hydrops in eyes with KC.