Rose Bengal Electromagnetic Activation with Green Light for Infection Reduction (REAGIR) Study: A randomized, double-masked, sham-controlled clinical trial of rose bengal photodynamic therapy.

Objective: To determine the effect of adjunctive rose-bengal photodynamic therapy (RB-PDT) in the treatment of fungal, Acanthamoeba, and smear/culture negative infectious keratitis.

Study Design: This international, randomized, double-masked, sham controlled clinical trial, randomizes patients with corneal ulcers in a 1:1 fashion to one of two treatment arms: 1) Topical antimicrobial plus sham RB-PDT or 2) Topical antimicrobial plus RB-PDT.

Main Outcome Measures: The primary outcome was best spectacle-corrected visual acuity (BSCVA) at 6 months. Secondary outcomes included BSCVA at 3 weeks and 3 months, infiltrate and/or scar size at 3 weeks, 3 months, and 6 months, corneal perforation (CP) and/or the rate of therapeutic penetrating keratoplasty (TPK), microbiological cure rate.

Results: A total of 330 patients were enrolled. Isolated organisms included filamentous fungus (N=301; 91%), Acanthamoeba (N=10; 3%) and culture/smear negative (N=19; 6%). There was no evidence of a benefit of RB-PDT versus sham for BSCVA at 6 months (-0.0004; 95% CI -0.13 to 0.13; P=0.62). There was evidence of a statistically significant interaction between treatment arm and fungal organism with regards to BSCVA (P=0.02). RB-PDT improved BSCVA among eyes with Fusarium infections (difference = -0.17 logMAR, 95% CI -0.37 to 0.03) but worsened BSCVA among eyes with Aspergillus infections (difference = 0.39 95% CI -0.03 to 0.80 logMAR). Scar size was 0.45 mm smaller in the RB-PDT group at 3 weeks (95% CI -0.76 to -0.15; P=0.004) but this finding was no longer statistically significant at 3 and 6 months. There was no difference between groups in rates of CP/TPK or microbiological cure.

Conclusions: We were unable to find a benefit of adjuvant RB-PDT overall. Future directions may include trying alternative photosensitizers and/or other treatment algorithms.