Corneal Confocal Microscopy in Dry Eye Treated with Corticosteroids.

ABSTRACT

PURPOSE: To evaluate, by in vivo laser scanning confocal microscopy (LSCM), the corneal findings in moderate-to-severe dry eye patients before and after treatment with topical corticosteroid and to associate the confocal findings to the clinical response. METHODS: Fifty eyes of 50 patients with moderate-to-severe dry eye were included in this open-label, masked study. Exclusion criteria were any systemic or ocular condition (other than dry eye) and any systemic or topical treatment (except artificial tears), ongoing or performed in the previous 3 months, with known effect on the ocular surface. All patients were treated with loteprednol etabonate ophthalmic suspension 0.5% qid for 4 weeks. Baseline and follow-up (day 30 ± 2) visits included Ocular Surface Disease Index (OSDI) questionnaire, full eye examination, and central cornea LSCM. We compared data obtained before and after treatment and looked for associations between baseline data and steroid-induced changes. Based on the previously validated OSDI Minimal Clinically Important Difference, we reanalyzed the baseline findings comparing those patients clinically improved after steroids to patients not clinically improved after steroids. RESULTS: Ocular Surface Disease Index score and LSCM dendritic cell density (DCD) significantly decreased after treatment. Baseline DCD correlated with both OSDI and DCD steroid-related changes (r = -0.44, p < 0.05 and r = -0.70, p < 0.01, respectively; Spearman) and was significantly higher in patients clinically improved after steroids than in patients not clinically improved after steroids (164.1 ± 109.2 vs. 72.4 ± 45.5 cells/mm2, p < 0.01; independent samples t test). CONCLUSIONS: Laser scanning confocal microscopy examination of DCD allows detection of treatment-related inflammation changes and shows previously unknown associations between confocal finding and symptoms improvement after treatment. These promising preliminary data suggest the need for future studies testing the predictive value of DCD for a clinical response to topical corticosteroids.