First 100: learning curve for Descemet stripping automated endothelial keratoplasty

Endothelial keratoplasty is largely replacing penetrating keratoplasty for the routine treatment of corneal endothelial failure. The aim of the study was to describe the performance, complications, and outcome of the first 100 Descemet stripping automated endothelial keratoplasty (DSAEK) procedures performed at a major medical center, with an emphasis on the learning curve.

A retrospective, comparative case series study was conducted by a single surgeon at a tertiary, university-affiliated medical center. Data were collected on 100 consecutive DSAEK procedures performed between September 2008, when the technique was introduced in the ophthalmology department, and January 2011. Main outcome measures include best-corrected visual acuity, graft dislocation rate, primary failure rate, and endothelial cell loss. Findings were compared between the first (early group) and last (late group) 50 cases.

Mean (SD) logMAR best-corrected visual acuity at 1 year improved from 1.02 ± 0.36 to 
0.38 ± 0.35 in the early group (p<0.05) and from 0.93 ± 0.48 to 0.25 ± 0.21 in the late group (p<0.05) 
(n = 86). Although the late group included more complex cases (presence of anterior chamber intraocular lens, need for combination cataract surgery or secondary scleral intraocular lens fixation), graft dislocation was more common in the early group (20% versus 10%, p = 0.26). Primary and late graft failures occurred only in the early group (12% vs 0, p = 0.03).

Rates of primary failure and disc dislocation for DSAEK decrease as surgeons gain experience with the procedure, and the number of functional grafts increases accordingly. Visual outcome improves regardless of surgical experience.