Graft outcomes, reinterventions, and macular morbidity after DMEK in vitrectomized versus non-vitrectomized eyes

To compare visual, anatomic, and graft-related outcomes after Descemet membrane endothelial keratoplasty (DMEK) in vitrectomized versus non-vitrectomized eyes, and to identify factors associated with adverse outcomes in these eyes.

This retrospective study analyzed DMEK performed between January 2018 and January 2025. Fifty-four vitrectomized eyes with pseudophakic bullous keratopathy (PBK) were compared with 50 non-vitrectomized PBK controls without major ocular comorbidity. Outcomes included best-corrected visual acuity (BCVA), central corneal thickness (CCT), endothelial cell density (ECD) and ECD loss, graft-related reinterventions, and cystoid macular edema (CME).

104 eyes were analyzed. BCVA and CCT improved significantly in both groups (p < 0.001). At 12 months, vitrectomized eyes had lower ECD (1458 ± 545 vs 1759 ± 436 cells/mm²; p = 0.030) and greater ECD loss (50.8 ± 18.6% vs 38.5 ± 15.2%; p = 0.031). CME (35.2% vs 6.0%; p < 0.001) and repeat DMEK (20.4% vs 4.0%; p = 0.016) were more frequent in vitrectomized eyes. Rebubbling (29.6% vs 16.0%) and penetrating keratoplasty (7.4% vs 0%) were more common but nonsignificant (p = 0.11, 0.12). A longer vitrectomy-to-DMEK interval was associated with CME (p = 0.013).

DMEK provides meaningful visual and anatomic improvement in vitrectomized eyes. However, versus controls they had greater endothelial cell loss, more frequent repeat DMEK, and markedly higher CME, whereas rebubbling and penetrating keratoplasty differences were nonsignificant. Given their heterogeneity, frequent capsular-zonular compromise, and the retrospective design, these associations warrant caution. Vitrectomized eyes represent a complex, higher-risk DMEK group warranting closer follow-up and individualized perioperative management.