Abstract Purpose To evaluate the surgery‐induced changes of astigmatism after Descemet membrane endothelial keratoplasty (DMEK) in eyes with failed previous penetrating keratoplasty (PK). Design Retrospective, interventional cohort study based on prospective DMEK database. Methods Evaluation of 15 eyes after DMEK surgery for endothelial decompensation after PK. Astigmatic changes before decompensation and after DMEK surgery were compared using anterior and posterior keratometry and True Net Power (TNP) in corneal tomography (Pentacam, Oculus). The differences in magnitude of astigmatism and axis of steep meridian were calculated by algebraic and vectorial methods, and significant differences to specified thresholds were determined by the non‐parametric Wilcoxon rank test. Results Surgically induced changes (SIC) in magnitude of astigmatism at anterior surface, posterior surface and TNP were −1.4dpt ± 3.7, −0.2dpt ± 0.8 and −1.3dpt ± 3.9, respectively. No significant changes in astigmatism and axis difference by algebraic method in comparison with a hypothetical median of 0dpt and 10° were observed. Calculation of astigmatism difference by vectorial method showed a significant deviation in anterior surface (observed median = 4.4dpt, p < 0.001), posterior surface (observed median 1.0dpt, p = 0.012) and TNP (observed median = 3.6dpt, p < 0.001). Conclusion Even if no significant change in magnitude and axis of astigmatism to acceptable thresholds could be shown after DMEK in failed PK using the algebraic method, there is a significant and relevant change in magnitude of astigmatism using the vectorial calculation. Based on these results, caution is advised when using toric intraocular lenses in PK eyes, as astigmatic change is expected when performing DMEK on eyes with failed PK.
Steinberg et al. (2026) studied this question.