Effects of corneal nerve density on the response to treatment in dry eye disease

To evaluate whether levels of corneal subbasal nerve fiber length (SNFL) in dry eye disease (DED) could prognosticate the level of improvement in signs and symptoms after treatment.Phase IV, double-masked, randomized clinical trial.Sixty patients with meibomian gland dysfunction-associated DED and 27 age-matched controls.Patients with DED were randomized to receive topical artificial tears, loteprednol etabonate 0.5%, or loteprednol etabonate 0.5%/tobramycin 0.3% twice daily for 4 weeks. At baseline, in vivo confocal microscopy of central cornea was performed in both eyes. Patients with DED were divided into 2 subgroups: those with low baseline SNFL and those with near-normal baseline SNFL for this purpose (the cutoff point: the mean SNFL in controls minus 2 standard deviations). Clinical signs and symptoms at baseline and after 4 weeks of treatment were compared between the subgroups with low and near-normal SNFL for all therapeutic groups.Symptom questionnaires, corneal fluorescein staining (CFS), conjunctival staining with lissamine green, tear break-up time, Schirmer's test, and SNFL.In patients with DED, baseline SNFL (17.06±5.78 mm/mm(2)) was significantly lower than in controls (23.68±3.42 mm/mm(2), P = 0.001). In the artificial tear and loteprednol groups, although no significant improvement in any sign or symptom was noted in patients with low baseline SNFL (16.84 mm/mm(2)), subjects with near-normal baseline SNFL (≥16.84 mm/mm(2)) showed significant improvement in both symptoms and CFS score (all P0.05). In the loteprednol/tobramycin group, no significant change was evident for any sign or symptom in either subgroup of low or near-normal baseline SNFL.Significant improvements in CFS and patient symptomatology after DED treatment were evident only in the subgroup with near-normal corneal SNFL. Consideration of SNFL may assist in explaining the variability of patients' response to DED therapy.