Greater risk of cystoid macular edema may exist after DMEK compared to DSAEK
Descemet membrane endothelial keratoplasty (DMEK) independently correlates with a greater risk of cystoid macular edema compared to Descemet stripping automated endothelial keratoplasty (DSAEK). Age and diabetes also are risk factors.
This single-center, retrospective, cohort study queried endothelial keratoplasties performed 2005–2018 (1909 DSAEK patients) and 2014–2018 (324 DMEK patients) with at least 18 months of follow-up. One eye was analyzed from patients who underwent bilateral surgery; eyes with preexisting retinal comorbidities, including diabetic macular edema, were excluded as were eyes that underwent endothelial keratoplasty during a surgeon’s learning curve. Patients suspected of having cystoid macular edema (reduced visual acuity of <20/30 despite a clear cornea and no discernible funduscopic findings on exam) underwent OCT (Spectralis HRA+OCT, Heidelberg, Germany). Cystoid macular edema was defined as intraretinal fluid spaces with or without subretinal fluid in the foveal region.
Cystoid macular edema was identified in 2.8% of the 2233 cases (2.4% in DSAEK eyes vs 5.6% in DMEK eyes, P = 0.001). Diabetes mellitus (OR = 3.16, P < 0.001) and DMEK (OR = 2.42, P = 0.003) were identified as independent risk factors for postoperative cystoid macular edema following endothelial keratoplasty. Using the cutoff identified by receiver operator curve analysis, subjects more than 67 years of age were also observed to be more likely to develop cystoid macular edema (OR = 2.35, P = 0.005). The mean onset of cystoid macular edema was 4.3 ± 6.63 months (range of 1–34 months).
This study comprises a large sample size, but it is still limited by a retrospective design. Sampling bias may have under-detected cystoid macular edema because OCT was obtained based on clinical suspicion rather than on a routine protocol. Moreover, undetected diabetic macular edema may have confounded the results. Preoperative OCT was used to exclude patients with baseline diabetic macular edema, but it is possible that occult or subtle retinal changes were missed, leading to the mislabeling of these cases as postoperative cystoid macular edema rather than persistent diabetic macular edema. Diabetic severity also was not included in the multivariate regression analysis.
Older patients (>67 years), diabetes mellitus, and DMEK are independent risk factors for postoperative cystoid macular edema following endothelial keratoplasty.
Financial disclosures: Dr. Christopher Sales discloses a financial relationship with Network Medical Review (Patents/Royalty).