Effects of Topical Insulin on Corneal Epithelium Healing After Corneal Crosslinking in Patients With Keratoconus

Patients from the Keratoconus Clinic of the Hospital de Clínicas, State University of Campinas (UNICAMP), Brazil, with a diagnosis of keratoconus and an indication for corneal crosslinking were evaluated with visual acuity, refraction, slit-lamp biomicroscopy, Goldmann applanation tonometry, corneal specular microscopy (EM-3000, Tomey) for endothelial cell count, and corneal tomography (Pentacam HR, Oculus).

Eligible patients who signed the informed consent form (with the legal guardian also signing for participants younger than 18 years) underwent crosslinking. After the procedure, participants were allocated to two groups (insulin and placebo) by block randomization with blocks of four, stratified by sex. The random sequence was generated at random.org/lists and concealed in sealed opaque envelopes prepared by an independent assistant. Envelopes were opened sequentially at enrollment by investigators who did not take part in outcome assessment. One participant underwent crosslinking in both eyes, which were allocated to different groups; the eye was the unit of analysis.

Crosslinking was standardized for all participants and performed by third-year ophthalmology residents supervised by attending physicians. Central 8.0-mm epithelial debridement was performed with an optical zone marker of the same diameter: 30% alcohol was applied within the marker for 10 seconds and dried with an ophthalmic sponge, and the epithelium was removed with a corneal scarifier spatula. Riboflavin 0.1% (400 mOsm) in a hydroxypropyl methylcellulose vehicle was used, with UVA irradiation at 9 mW/cm² for 10 minutes (fluence 5.4 J/cm²) using the Eyelink Crosslink device (Eyetec, São Carlos, Brazil). At the end of the procedure, the size and shape of the debrided area were checked with the 8.0-mm marker, and participants whose debridement did not meet this standard were excluded.

The insulin group received NPH insulin 50 IU/mL eye drops (Eye Pharma, São Paulo, Brazil) four times daily, and the placebo group received 0.5% methylcellulose eye drops (Eye Pharma), the vehicle of the insulin formulation, four times daily, until complete epithelial healing. The bottles were identical and were labeled on the prescription as "compounded eye drops". Participants and outcome assessors were masked to group allocation. Both groups received moxifloxacin 5 mg/mL every 6 hours (Oftalmox; Geolab), prednisolone acetate 10 mg/mL every 6 hours (Ster; União Química), 0.15% sodium hyaluronate every 4 hours (Hyabak; Laboratoires Théa) and a silicone hydrogel bandage contact lens (Air Optix plus HydraGlyde; Alcon) until complete epithelial healing. Moxifloxacin and the study eye drops were then discontinued, and prednisolone was tapered weekly until one month after the procedure.

Crosslinking was always performed on Tuesdays, and assessments took place 18, 42 and 66 hours after the procedure (postoperative days 1 to 3). If epithelialization was incomplete at 66 hours, a further visit was scheduled on postoperative day 6 and every 48 hours thereafter until complete epithelialization. At each visit, the epithelial defect was photographed with the Keratograph 5M (OCULUS) and slit-lamp biomicroscopy was performed. The defect area was measured with ImageJ software by an investigator masked to allocation and reviewed independently by a second masked investigator. Complete epithelial healing was defined as the absence of an epithelial defect after fluorescein instillation on Keratograph 5M images, confirmed by two examiners on slit-lamp examination.

The primary outcome is the corneal epithelialization rate (mm²/h). Secondary outcomes are time to complete epithelialization and adverse events, in addition to endothelial cell density. Later follow-up visits took place at 30 and 60 days after crosslinking.