Anterior segment optical coherence tomography (AS-OCT) assisted analysis of characteristics of graft dehiscence after Descemet membrane endothelial keratoplasty (DMEK) for failed penetrating keratoplasty.

Purpose: To analyse anterior segment optical coherence tomography (AS-OCT) parameters of graft dehiscence after Descemet membrane endothelial keratoplasty (DMEK) for graft failure post penetrating keratoplasty (PK).

Methods: Retrospective evaluation of AS-OCT images of 142 dehiscences post-DMEK in 75 eyes. Dehiscences' size, depth, location, correlation with graft-host interface (GHI) override and step at GHI were assessed.

Results: The majority of patients were male (58.7%) and mean age was 67 ± 13.6 years. Multiple dehiscences were observed in 49.3% eyes. Rebubbling was required once in 72% and multiple times in 24% eyes. Among 142 dehiscences, crossing over GHI was noted in 53.5%. The median distance of peripheral edge of dehiscences from GHI was 0.21 mm. Steps at GHI were noted in 41.5% where 22.5% dehiscences with step crossed the GHI. For dehiscences crossing versus not crossing GHI, no significant difference was noted in median depth (p = 0.268) and size (p = 0.206). For dehiscences crossing over GHI with presence versus absence of step, median depth (p = 0.23) and size (p = 0.196) showed no significant difference. No significant difference was noted in dehiscences' median depth (p = 0.16) and size (p = 0.926) among different step sizes. Incidence of dehiscence with or without steps (p = 0.8853) and graded as per size of steps, showed no significant difference.

Conclusion: Size and depth of dehiscence were not influenced by their crossing over GHI or the presence and size of steps. Dehiscences typically occurred in close vicinity to GHI, suggesting that DMEK graft should be placed 0.25 mm centrally from the prior PK interface, independent of GHI irregularities.