LAUSR

PURPOSE To compare the outcomes of Descemet stripping automated endothelial keratoplasty (DSAEK) with Descemet membrane endothelial keratoplasty (DMEK) for the treatment of failed penetrating keratoplasty (PKP). METHODS This is a retrospective chart review of patients with failed PKP who underwent DMEK or DSAEK. The median follow-up time for both groups was 28 months (range 6-116 months). Data collection included demographic characteristics, number of previous corneal transplants, previous glaucoma surgeries, best-corrected visual acuity, endothelial cell density, graft detachment and rebubble rate, rejection episodes, and graft failure. RESULTS Twenty-eight eyes in the DMEK group and 24 eyes in the DSAEK group were included in the analysis. Forty-three percent of eyes in the DMEK group and 50% of eyes in the DSAEK group had to be regrafted because of failure (P = 0.80). The most common reason for failure was persistent graft detachment (58%) in the DMEK group and secondary failure (58%) in the DSAEK group; hence, the time between endothelial keratoplasty and graft failure differed significantly between the groups (P = 0.02). Six eyes (21%) in the DMEK group and 7 eyes (29%) in the DSAEK group developed graft rejection (P = 0.39). Rejection was the cause of failure in 67% and 71% in the DMEK and DSAEK groups, respectively. The best-corrected visual acuity 6 months after surgery was better in the DMEK group compared with the DSAEK group (P = 0.051). CONCLUSIONS Both DSAEK and DMEK have a role in treating PKP failure. Primary failure due to persistent graft detachment was significantly higher in the DMEK group, although the overall failure rate in the medium term was similar.