Endothelial damage with two DSAEK insertion techniques performed by a novice corneal surgeon in residency training: a...

The aim was to determine which of 2 commonly used nonfold donor insertion techniques was advantageous for initial Descemet stripping automated endothelial keratoplasty (DSAEK) cases.

This involved an ex vivo, prospective comparative case series of 20 randomized DSAEK lenticule insertions. DSAEK insertions were performed by a single novice corneal surgeon (PGY4 resident) in human cadaver eyes. Ten grafts were inserted using a Sheet glide (Surgical Glide; Beaver-Visitec International Inc, Waltham, MA) and 10 were inserted using an inserter device (EndoSerter; Ocular Systems Inc, Winston-Salem, NC). The grafts were explanted, stained with trypan blue and alizarin red S, and photographed for comparison with 5 control grafts. Endothelial damage was quantitatively evaluated using Adobe Photoshop 10.0 CS3 software (Adobe Systems, San Jose, CA).

Endothelial cell loss (ECL) was 7.10% ± 2.27% in controls, 12.31% ± 4.74% in the inserter group, and 13.31% ± 5.46% in the Sheet glide group (P = 0.07). Early cases (cases 1-5) had a greater ECL compared with what later cases had (cases 6-10) for the Sheet glide group. This difference was significant for the Sheet glide group (40.42% reduction for cases 6-10, P = 0.04) but not for the EndoSerter group (32.5% reduction for cases 6-10, P = 0.11).

Quantitative analysis revealed no statistically significant difference in the ECL between the 2 methods. With surgeon experience, there was a trend toward less ECL using both methods but especially with the Sheet glide. A novice corneal surgeon may effectively use either of these nonfold methods for initial cases. The cadaver eye model described may be a potentially useful wet-laboratory tool for novice surgeons to practice DSAEK lenticule insertion.