Purpose: Descemet membrane endothelial keratoplasty (DMEK) has emerged as a treatment for corneal endothelial failure after penetrating keratoplasty (PKP); however, outcomes in this high-risk population remain limited. This meta-analysis evaluates the efficacy and safety of DMEK in eyes with prior PKP failure. Methods: We conducted a systematic review and meta-analysis, following PRISMA guidelines. PubMed, Web of Science, and CENTRAL databases were searched from inception to February 2025. Meta-analyses were conducted using a random-effects model. Meta-analyses of weighted means with 95% CIs and logit-transformed proportions with 95% CIs were performed to assess differences in best-corrected visual acuity (BCVA), endothelial cell density (ECD), and central corneal thickness (CCT) post-PKP failure compared with post-DMEK. Results: We included 14 studies (n = 386 eyes). The weighted mean age for the patients was 65.6 years (95% CI: 58.5–72.7), with a slight male predominance (55.1% 95% CI: 49.7%–60.5%). Keratoconus (35.8% 95% CI: 27.5%–45.1%) and Fuchs endothelial dystrophy (23.1% 95% CI: 16.9%–30.7%) were the most common indications for initial PKP. Patients had undergone an average of 1.4 prior PKPs (95% CI: 0.79–2.0), with a mean interval of 12.1 years (95% CI: 5.7–18.6) since their last PKP. Graft attachment was achieved in 62.7% (95% CI: 52.9%–71.5%) of cases. Complications included detachment (38.4% 95% CI: 29.4%–48.3%) and rebubbling (37.9% 95% CI: 28.9%–47.9%). DMEK significantly improved BCVA (weighted mean difference: −0.62 logMAR 95% CI: −0.76 to −0.48, P < 0.001). ECD declined by 1046.7 cells/mm 2 (95% CI: −1346.1 to −747.2, P < 0.001), whereas CCT decreased by 192.2 µm (95% CI: −239.1 to −145.4, P < 0.001). Conclusions: DMEK is a viable option for failed PKP, providing visual rehabilitation and corneal clearance, yet with relatively considerable endothelial cell loss. This highlights the need for new techniques to improve endothelial survival and reduce postoperative complications.
Qedair et al. (Sun,) studied this question.