Descemet membrane endothelial keratoplasty compared with ultrathin Descemet stripping automated endothelial keratoplasty: a meta-analysis

Aims This study aims to compare the clinical outcome of Descemet membrane endothelial keratoplasty (DMEK) and ultrathin Descemet stripping automated endothelial keratoplasty (UT-DSAEK) in patients with corneal endothelial dysfunction due to Fuchs' endothelial dystrophy or pseudophakic bullous keratopathy.Methods We conducted a meta-analysis using a literature search of Embase, PubMed, Cochrane CENTRAL, ClinicalTrials.gov and WHO ICTRP databases. We included randomised controlled trials (RCTs) and cohort studies that compared DMEK and UT-DSAEK (graft<130 mu m), with a follow-up of >= 12 months, published until 20 February 2022. We used the Revised Cochrane risk-of-bias tool for RCTs and the Risk of Bias in Non-Randomised Studies-of Interventions system for cohort studies.Results Out of 144 records, 8 studies (3 RCTs, 2 fellow-eye studies and 3 cohort studies) were included, encompassing 376 eyes, (N=187 DMEK vs N=189 UT-DSAEK). The 12-month logarithm of the minimum angle of resolution best-corrected visual acuity (BCVA) was better post-DMEK (mean difference -0.06 (95% CI -0.10 to -0.02)), but with higher rebubbling risk: OR 2.76 (95% CI 1.46 to 5.22). Heterogeneity was significant I-2=57%. Findings were consistent when excluding retrospective studies, including only studies with low risk of bias or RCTs only. An analysis of studies with mean DSAEK grafts <70 mu m showed no significant difference in BCVA between the procedures. Publication bias was found in the BCVA analysis (Egger's test p=0.023).Conclusions Post-DMEK BCVA is superior to post-UT-DSAEK when using <130 mu m grafts. DSAEK grafts <70 mu m may not significantly differ from DMEK. The higher risk of rebubbling with DMEK necessitates an appropriate selection of patients.