Descemet's stripping automated endothelial keratoplasty: the tenuous relationship between donor thickness and postoperative vision

Abstract

Purpose: To determine whether preoperative donor thickness has a relationship with postoperative visual acuity after Descemet's stripping automated endothelial keratoplasty (DSAEK).

Design: Retrospective correlation and comparative analysis of an interventional case series.

Participants: A total of 418 eyes of 292 patients undergoing DSAEK surgery for Fuchs' endothelial dystrophy without visual loss from comorbidities.

Methods: Descemet's stripping automated endothelial keratoplasty was performed in 548 eyes with Fuchs' dystrophy, and preoperative graft thickness (GT) was recorded. After exclusion of patients with confounding variables that would affect postoperative visual acuity, postoperative best spectacle-corrected visual acuity (BSCVA) was measured at 6 months in 418 eyes. Pearson's correlation analysis was performed between preoperative GT and BSCVA. Cases were split into deciles on the basis of GT and BSCVA and then compared with 1-way analysis of variance (ANOVA) and chi-square test.

Main outcome measures: Best spectacle-corrected visual acuity at 6 months postoperatively.

Results: Mean GT of the series was 162.9±29.0 μm (range, 80-265 μm), and mean Snellen BSCVA was 20/28 with a range of 20/16 to 20/70. There was a weak correlation between GT and BSCVA that was significant (R = 0.236, P<0.001) but only accounted for 5% of the visual outcome (R(2) = 0.056). Visual outcome was best within the thinnest decile group of 45 donors (GT range, 80-124), with a mean Snellen BSCVA of 20/25 (range, 20/20-20/50), and worst within the thickest decile group of 41 donors (GT range, 200-265), with a mean Snellen BSCVA of 20/33 (range, 20/20-20/70). Post hoc comparison of BSCVA between the thickest and thinnest groups was significant (P = 0.006).

Conclusions: Preoperative GT may have a small effect on visual outcome in the extremes of thickness, but not in the common range of 100 to 200 μm. Donor thickness has a tenuous relationship with visual outcome, accounting for only 5% of the variance in vision between patients, and should play a minimal role in surgical planning.