OBJECTIVE. UV-A/Riboflavin crosslinking of corneal collagen fibers (CXL) is a highly promising therapy for corneal melting in humans. A prospective interventional, non-randomized, controlled study was conducted to compare the stabilizing effect of CXL treatment on melting keratitis in dogs and cats and the complication rate of CXL to those of standardized intensive medical treatment. PROCEDURES. Forty-nine eyes with melting keratitis were included in the study between October 2009 and October 2012. All eyes were treated according to the same medical treatment protocol. Nineteen eyes were CXL-treated and 30 eyes were not. Follow-up included slit-lamp examination, fluorescein staining, ulcer size measurement, stromal stability evaluation, photographic documentation and documentation of complications. RESULTS. Five of 19 eyes in the CXL group and 9/30 eyes in the control group required rescue stabilization due to continued melting. Seven of the 9 control group corneas stabilized after rescue CXL treatment. At initial presentation, the ulcers in the canine CXL group were significantly deeper and larger than in the control group. Ulcer deepening during follow-up was more pronounced in the canine control group than in the canine CXL group. CXL treatment related complications were not observed. CONCLUSIONS. Based on the similar failure rates in the control and CXL treatment groups despite the poorer initial situation in the CXL group, the tendency for the ulcers in the control group to deepen and the stabilization of all corneas receiving CXL rescue treatment, we believe that CXL has its place as an adjunctive therapy for melting keratitis in veterinary ophthalmology. RESULTS.Five of 19 eyes in the CXL group and 9/30 eyes in the control group required 39 rescue stabilization due to continued melting. Seven of the 9 control group corneas stabilized 40 after rescue CXL treatment. At initial presentation, the ulcers in the canine CXL group were 41 significantly deeper and larger than in the control group. Ulcer deepening during follow-up 42 was more pronounced in the canine control group than in the canine CXL group. CXL 43 treatment related complications were not observed. 44 CONCLUSIONS. Based on the similar failure rates in the control and CXL treatment groups 45 despite the poorer initial situation in the CXL group, the tendency for the ulcers in the control 46 group to deepen and the stabilization of all corneas receiving CXL rescue treatment, we 47 believe that CXL has its place as an adjunctive therapy for melting keratitis in veterinary 48 ophthalmology. 49

Objective Corneal collagen cross-linking with riboflavin and UV-A (CXL) decreases corneal oedema and increases visual acuity in human patients with bullous keratopathy. Presumed mechanisms are an increase in collagen packing density and a reduction in stromal swelling pressure. We present two cases in which CXL was used to treat bullous keratopathy in dogs. Procedures Four eyes of two dogs with painful bullous keratopathy-induced corneal erosions that were resistant to prior therapy were treated with CXL. Both corneas of the second patient were dehydrated to AE 400 lm corneal thickness using topical 70% glycerol solution immediately prior to CXL. Follow-up included slit-lamp examination, fluorescein staining and photographic documentation in both cases and highresolution ultrasound examination in the second patient. Results All four eyes were comfortable and fluorescein negative at 1-week post-CXL and remained so for the rest of the follow-up period (17.5 months for case 1 and 6 months for case 2). The owner of the first patient reported a less oedematous cornea and improvement in vision that lasted for 6 months. Despite a reported lack of improvement in vision in the second patient, corneal thickness initially decreased, but was back at baseline thickness at the 4-month recheck. Conclusions Similar to humans, CXL might become a useful treatment option for bullous keratopathy-induced therapy-resistant corneal erosions in dogs. Patient comfort was greatly improved, but corneal thickness decrease was not as long-lasting as reported for humans. The presently used protocols might need modification to fit the dog cornea.