Prediction of Descemet Membrane Perforation During Deep Anterior Lamellar Keratoplasty in Patients With Keratoconus With Stromal Scar

Abstract

Objectives: To report big-bubble deep anterior lamellar keratoplasty (DALK) in patients with keratoconus with corneal stromal scar, and to investigate factors that can influence intraoperative Descemet membrane (DM) perforation.

Methods: A retrospective study included patients with keratoconus with central stromal scar that underwent DALK using the big-bubble technique. Best spectacle-corrected visual acuity (BSCVA), keratometry, minimum corneal thickness (MCT), stromal scar depth, and endothelial cell density (ECD) were recorded. Receiver operating characteristic (ROC) curves were analyzed to predict an intraoperative perforation related to stromal scar.

Results: Thirty-eight eyes of 38 patients with keratoconus (mean age: 30.0±9.2 years) were included in this study. Thirty-four eyes had successful DALK; in four eyes, the procedure was converted to penetrating keratoplasty because of DM macroperforations. Successful big-bubble formation was achieved in 24 of 38 (63.2%) eyes, whereas in 10 eyes, manual dissection was used to complete the DALK. Mean preoperative and postoperative second year BSCVA were 1.3 (±0.4) and 0.5 (±0.2) logarithm of the minimum angle of resolution (P<0.001), keratometry values were 65.5±7.4 and 42.8±7.8 diopters (P<0.001), and ECD were 2,740±359 and 2,279±452 cells/mm (P<0.001), respectively. Using ROC analysis, the area under curve value to predict DM perforation was found to be 0.792 (scar/MCT ratio, P=0.04) and the best cutoff point for scar depth/MCT ratio was 53%.

Conclusion: Big-bubble DALK is effective in patients with keratoconus with stromal scars. Scar depth/MCT ratio seems to predict the DM perforation.