Is Endothelial Keratoplasty An Option To Treat Endothelial Failure After Penetrating Keratoplasty In Eyes With Suboptimal Conditions For A New Full Thickness Graft?

Endothelial keratoplasty has become the procedure of choice to treat endothelial failure and may be an alternative to a new full thickness procedure in eyes with endothelial failure after a previous penetrating keratoplasty (PK). Our purpose it to evaluate the long-term outcomes of Descemet Stripping Automated Endothelial Keratoplasty (DSAEK), following PK with endothelial failure, in eyes with poor prognosis for a new PK.

Cornea Unit, Ophthalmology Department, Centro Hospitalar Universitário do Porto, Porto, Portugal

Retrospective evaluation of eyes submitted to DSAEK due to endothelial failure after PK, between 2014 and 2020. In all cases, a new PK would be associated with a poor prognosis due to associated ocular conditions, namely: 1) three or more previous failed grafts; 2) corneal neovascularization in at least three quadrants; 3) poor ocular surface conditions; 4) significant posterior segment pathology. Only eyes with at least two years of follow-up were included. Best corrected visual acuity (BCVA) was evaluated with the logMAR scale. Both BCVA and graft survival were evaluated one, two, three and five years after surgery. The number of PK performed before DSAEK and the number of eyes requiring re-bubbling or additional grafts were recorded.

Thirty-five eyes of 35 patients were included. Mean follow-up time was 3.9±2.3 [2-9] years. The mean number of PK before DSAEK was 1.8±1.0 [1-4]. Graft survival rate was 79% at year 1, 71% at year 2, 60% at year 3 and 46% at year 5.  BCVA improved from 1.6±0.5 to 1.4±0.7 logMAR (p=0.016), one month after DSAEK. In eyes with a functioning graft, there was further improvement 1 year after surgery (1.3±0.7 vs 1.0±0.8 logMAR, p=0.005), without further changes at year 2 (1.0±0.8 vs 1.1±1.7 logMAR, p=0.405), year 3 (0.7±0.2 vs 0.8±0.3 logMAR, p=0.072) and year 5 (0.7±0.3 vs 0.8±0.4 logMAR, p=0.580). Re-bubbling due to graft detachment was needed in 17%. Nine eyes (26%) required at least one more keratoplasty, after a mean time of 1.7±1.7 years.

Despite the associated ocular conditions and comorbidities of the eyes included in this study being associated with a poor prognosis, DSAEK allowed reasonable long-term graft survival rates and acceptable visual outcomes. Thus, DSAEK seems to be a good option for endothelial failure post PK in these special cases.