Descemet Stripping Automated Endothelial Keratoplasty for Failed Penetrating Keratoplasty: Influence of the Graft-Host Junction on the Graft Survival Rate.

PURPOSE

To investigate the clinical results of Descemet stripping automated endothelial keratoplasty (DSAEK) for failed penetrating keratoplasty (PK) and the influence of the graft-host junction (GHJ) on the graft survival rate.

METHODS

Data were retrospectively collected on patient demographics, visual outcomes, complications, and graft survival rate for 17 eyes of 16 patients who underwent DSAEK for failed PK. The graft survival rate was compared between the eyes when divided into a bump group and a well-aligned group according to the shape of the GHJ detected on anterior segment optical coherence tomography.

RESULTS

The most common indication for initial PK was bullous keratopathy after glaucoma surgery (35.3%). Seven eyes (41.2%) were classified into the bump group and 10 eyes (58.8%) into the well-aligned group. The mean best-ever documented visual acuity (BDVA) after DSAEK was 0.33 logMAR. Postoperatively, almost 70% of eyes achieved a BDVA that was within 0.2 logMAR of their preoperative BDVA. Graft detachment occurred in 29.4% of eyes and primary graft failure in 17.6%. All primary failures occurred in the bump group. The cumulative graft survival rate was 82.3% at 1 year, 73.2% at 2 years, and 58.6% at 3 years. Graft failure was more likely in eyes in the bump group than in those in the well-aligned group (P = 0.037, Wilcoxon test).

CONCLUSIONS

DSAEK for failed PK had a favorable outcome in this study. However, the GHJ should be assessed carefully before performing the procedure.