Primary graft failure in preloaded versus surgeon prepared DSAEK: a clinical and ex vivo correlation study.

Purpose: To evaluate whether primary graft failure (PGF) rates and endothelial cell loss (ECL) differ between surgeon-trephined/loaded and eye bank-preloaded Descemet stripping automated endothelial keratoplasty (DSAEK) grafts.

Setting: Tertiary care academic center.

Design: Retrospective case series and ex vivo laboratory study.

Methods: Single-center review of 69 eyes undergoing DSAEK (2020-2023) divided into three groups based on graft preparation and eye bank source: A) eye bank-preloaded in EndoSerter (PL-ES), B) eye bank-preloaded in glass tube (PL-GT), and C) surgeon-trephined/loaded in EndoSerter (SL-ES). The primary clinical outcome was PGF (persistent corneal edema >3 months despite medical therapy). Secondary outcomes included visual outcomes and clinical complications (cystoid macular edema (CME) and rebubbling rate (RR)). The primary outcome of the laboratory study was percentage ECL (%ECL).

Results: All lenticules were ≤80 µm. PGF was significantly higher in the PL-ES group (8/20; 40%) and PL-GT group (4/7; 57%) compared to the SL-ES group (1/42; 2%) (p < 0.0001). Grafts requiring rebubbling were more likely to experience PGF (p = 0.0176). PGF eyes were regrafted with SL-ES DSAEK; none experienced subsequent PGF. There were no significant differences in CME, RR, corrected distance visual acuity, and spherical equivalent. The ex vivo study included 9 PL-ES, 6 PL-GT, and 6 SL-ES grafts. %ECL was higher in both preloaded groups than SL-ES (p = 0.0354), though significance was not demonstrated on post-hoc analysis. Fluorescent imaging revealed more cell loss in preloaded grafts.

Conclusions: Preloading and storage of DSAEK grafts may impact PGF rates clinically and ECL ex vivo.