Objective To investigate the inhibitory effect of bromfenac sodium hydrate ophthalmic solution on corneal neovascularization (CNV) induced by alkali burn.Methods A total of 192 specific pathogen free (SPF) degree adult male Sprague-Dawley (SD) rats were used in this study.One hundred and seventy-two rats were chosen to establish CNV model with alkali burn in the right eyes.Following alkali burn,rats were randomly divided into CNV group,model control group,bromfenac sodium group and fluorometholone group,with 43 rats (43 eyes) in each group.Another 20 rats (40 eyes) served as normal control group.One day after modeling,the model control group,bromfenac sodium group and fluorometholone group received phosphate buffer saline (PBS),bromfenac sodium hydrate ophthalmic solution and 0.1％ fluorometholone eye drops,respectively.The state of cornea and anterior chamber and the growth of CNV of rats in each group were observed by slit-lamp microscope every day after modeling.At 1,3,7,14,21 and 28 days after modeling,the anterior segment photos of the experimental eyes were captured,and the percent of cornea areas covered by CNV was calculated.At 7,14 and 28 days after modeling,the eye tissue sections were stained with hematoxylin and eosin staining and immunohistochemistry staining to evaluate the expressions of CD45 and VEGF-A.Real-time quantitative PCR and enzyme-linked immunosorbent assay (ELISA)were used to detect the expression of COX-2 and VEGF mRNA and protein level.The use and care of the animals complied with the Statement of the Association for Research in Vision and Ophthalmology(ARVO).Results Each model group showed corneal edema and opacification 1 day after modeling.The corneal edema was aggravated 7 days after modeling.On the 14th day after modeling,the degree of corneal opacity and edema decreased gradually.On the 28th day after modeling,leucoma was observed in CNV group and model control group,and nebula was observed in bromfenac sodium group and fluorometholone group.At 7,14,21 and 28 days after modeling,the percentages of CNV areas in bromfenac sodium group and fluorometholone group were significantly lower than those in CNV group and model control group (all at P＜0.05).No significant difference was found in the percentage of CNV areas between bromfenac sodium group and fluorometholone group at various time points (all at P＞0.05).On the 7th day after modeling,the thinning of corneal epithelial layer,edema and arrangement disorder of stroma layer were observed,and the expression of VEGF-A was positive in all model groups;a small amount of CD45 positive inflammatory cell infiltrations were observed in CNV group and model control group.On the 14th and 28th day after modeling,CNV was seen in the center of cornea in CNV group and model control group;the epithelial keratosis and reduction of corneal edema were seen in each group,and no inflammatory cell infiltration was observed in each group.On the 7th day after modeling,the expressions of COX-2 and VEGF mRNA in CNV group and model control group were significantly higher than those in normal control group,bromfenac sodium group and fluorometholone group (all at P ＜ 0.05),the expressions of COX-2 and VEGF protein in bromfenac sodium group were significantly lower than those in CNV group (all at P＜0.05).The corneal peroration rate in model control group and bromfenac sodium group was 10％ (1 case in 10 rats).The corneal perforation rate in fluorometholone group was 30％ (3 cases in 10 rats).In each model group,10％ to 30％ rats had hyphema.Conclusions Bromfenac sodium hydrate ophthalmic solution can inhibit the formation and growth of CNV after alkali burn in rats.This effect may be mediated by regulating COX-2 expression,reducing inflammation and inhibiting VEGF production.