PURPOSE: It is unclear which patients unexpectedly have a hyperopic refractive outcome after combined Descemet membrane endothelial keratoplasty and cataract surgery (triple DMEK). We assessed how corneal shape predicts hyperopia after triple DMEK. DESIGN: Retrospective cohort study. METHODS: Patients with Fuchs endothelial corneal dystrophy (FECD) with Scheimpflug examinations before uncomplicated triple DMEK at a tertiary referral center were included. The arithmetic error was calculated (stable postoperative refraction minus predicted refraction). Using multinomial logistic regression, risk ratios of > +0.5 diopter (D) hyperopic and > 0.5 D myopic arithmetic errors were calculated. RESULTS: In 112 eyes, the median predicted refraction was -0.43 D (interquartile range [IQR], -0.47 to -0.17) with an achieved refraction of -0.63 to 0.56 (IQR). The arithmetic error was 0.34 D (IQR, -0.22 to 0.81). A hyperopic arithmetic error was present in 46% of eyes. FECD eyes with an oblate posterior cornea (Q value >0) had a 3.0 times higher risk of hyperopic shift after triple DMEK (95% confidence interval [CI], 1.3-7.0; P = .011), compared to spherical or prolate corneas (Q value ≤ 0). In eyes with posterior Q > 0, the mean prediction error was +0.50 D higher than in eyes with negative Q values (95% CI, 0.19-0.82; P = .002), independent of corneal thickness. CONCLUSIONS: Hyperopic surprises after triple DMEK particularly occur in corneas that are flatter centrally than the periphery because of edematous changes (oblate posterior profile). Eyes with a positive Q value on Scheimpflug imaging should be considered for additional power at the intraocular lens level.
PURPOSE: It is unclear which patients unexpectedly have a hyperopic refractive outcome after combined Descemet membrane endothelial keratoplasty and cataract surgery (triple DMEK). We assessed how corneal shape predicts hyperopia after triple DMEK. DESIGN: Retrospective cohort study. METHODS:Patients with Fuchs endothelial corneal dystrophy (FECD) with Scheimpflug examinations before uncomplicated triple DMEK at a tertiary referral center were included. The arithmetic error was calculated (stable postoperative refraction minus predicted refraction). Using multinomial logistic regression, risk ratios of > +0.5 diopter (D) hyperopic and > 0.5 D myopic arithmetic errors were calculated. RESULTS: In 112 eyes, the median predicted refraction was -0.43 D (interquartile range [IQR], -0.47 to -0.17) with an achieved refraction of -0.63 to 0.56 (IQR). The arithmetic error was 0.34 D (IQR, -0.22 to 0.81). A hyperopic arithmetic error was present in 46% of eyes. FECD eyes with an oblate posterior cornea (Q value >0) had a 3.0 times higher risk of hyperopic shift after triple DMEK (95% confidence interval [CI], 1.3-7.0; P = .011), compared to spherical or prolate corneas (Q value ≤ 0). In eyes with posterior Q > 0, the mean prediction error was +0.50 D higher than in eyes with negative Q values (95% CI, 0.19-0.82; P = .002), independent of corneal thickness. CONCLUSIONS: Hyperopic surprises after triple DMEK particularly occur in corneas that are flatter centrally than the periphery because of edematous changes (oblate posterior profile). Eyes with a positive Q value on Scheimpflug imaging should be considered for additional power at the intraocular lens level.