UCSF

Importance: Adjunctive topical corticosteroids and/or corneal cross-linking (CXL) have the potential to improve outcomes in bacterial keratitis. Objective: To determine the benefit of adjunctive topical difluprednate and CXL with riboflavin in addition to topical antibiotics. Design, Setting, and Participants: This was a National Institutes of Health (NIH)-funded, sham, placebo-controlled trial randomizing participants to topical moxifloxacin, 0.5%, plus topical placebo plus sham CXL, vs topical moxifloxacin, 0.5%, plus difluprednate, 0.05%, plus sham CXL, vs topical moxifloxacin, 0.5%, plus difluprednate, 0.05%, plus CXL. Between September 2020 and October 2023, participants in clinics at the Aravind Eye Hospitals in India and Bascom Palmer Eye Institute, University of Miami, in Miami, Florida, were screened for inclusion. Included participants had smear- and/or culture-positive bacterial corneal ulcers with Snellen visual acuity of 20/40 or worse. Main Outcomes and Measures: The primary outcome was logMAR best spectacle-corrected visual acuity (BSCVA) at 6 months. Results: Of the 1992 participants with smear-positive bacterial corneal ulcers screened, 280 participants (14%; mean [SD] age, 51 [16] years; 182 male [65%]) were enrolled. After controlling for baseline, there was no difference in 6-month VA with adjunctive topical steroids vs placebo (-0.04; 95% CI, -0.18 to 0.09; P = .58) or with adjunctive CXL and topical steroids vs topical steroids alone (0.04; 95% CI, -0.09 to 0.17; P = .62). Adjunctive topical corticosteroids also did not improve scar size at 6 months after controlling for baseline infiltrate and/or scar size (-0.22; 95% CI, -0.53 to 0.10; P = .65). CXL plus corticosteroids increased scar size compared with steroids alone after controlling for baseline infiltrate and/or scar size (0.56; 95% CI, 0.20-0.92; P = .02). There was 0.78 (95% CI, 0.27-2.24; P = .65) times the hazard of perforation or the need for therapeutic penetrating keratoplasty (TPK) in the early-steroid arm and 0.48 (95% CI, 0.14-1.67; P = .25) times the hazard of perforation or the need for TPK in the CXL arm after controlling for infiltrate depth. Conclusions and Relevance: Results of this randomized clinical trial reveal that adjunctive topical corticosteroids were not superior to placebo, and adjunctive CXL had increased scar size, suggesting that these alternative therapies, on average, may not be superior to topical antibiotics alone for treatment of bacterial keratitis. Trial Registration: ClinicalTrials.gov Identifier: NCT04097730.