Postoperative, but not preoperative, central corneal thickness correlates with the postoperative visual outcomes of Descemet membrane endothelial keratoplasty

Abstract

Descemet membrane endothelial keratoplasty (DMEK) restores visual acuity in patients with progressive corneal endothelial diseases such as Fuchs endothelial corneal dystrophy (FECD). However, patients often prefer to delay the surgery as long as possible, even though outcomes are poorer in advanced FECD. A recent study proposed that preoperative central corneal thickness (CCT) of ≥625 μm associated with worse best spectacle-corrected visual acuity (BSCVA) after DMEK for FECD. Since this threshold could signal to both surgeons and patients when to perform DMEK, we further explored the relationship between CCT and BSCVA with a retrospective cohort study. The cohort consisted of all patients with FECD who underwent DMEK in a tertiary-care hospital in 2015-2020 and were followed for 12 months. Extremely decompensated corneas were not included. Relationships between preoperative CCT and BSCVA on days 8 and 15 and months 1, 3, 6, and 12 were examined with Pearson correlation analyses. Eyes with preoperative CCT <625 or ≥625 μm were also compared in terms of postoperative BSCVA. Relationships between postoperative CCT and final BSCVA were also explored. The cohort consisted of 124 first-operated eyes. Preoperative CCT did not correlate with postoperative BSCVA at any timepoint. Eye subgroups did not differ in postoperative BSCVA. However, postoperative CCT at 1-12 months correlated significantly with 12-month BSCVA (r = 0.29-0.49, p = 0.020-0.001). Thus, postoperative, but not preoperative, CCT correlated with postoperative BSCVA. This phenomenon may reflect factors that distort preoperative CCT measurements but disappear after surgery. This observation and our analysis of the literature suggest that while there is a relationship between CCT and post-DMEK visual acuity, preoperative CCT measurements may not always adequately reflect that relationship and may therefore not be a reliable predictor of DMEK visual outcomes.