2010 OMIG, Abstract 3

OMIG Main Page | 2010 Abstracts | < Previous| Next >

The steroids for corneal ulcers trial (SCUT) baseline data and Bayesian statistic inquiry
C.W. See¹, N.R. Acharya¹, M. Srinivasan², M.E. Zegans³, K. Ray¹, D. Glidden⁴, C. Oldenburg¹, T. Porco^1,4, T.M. Lietman^1,4
¹ F.I. Proctor Foundation and the Dept. of Ophthalmology, University of California San Francisco, CA, ²Aravind Eye Care System, Madurai, India, ³ Dept. of Ophthalmology, Dartmouth Medical School, Lebanon, New Hampshire, DE, ⁴ Dept. of Epidemiology and Biostatistics, University of California, San Francisco, CA

Purpose: 1) To report on the baseline characteristics of the Steroids for Clinical Ulcers Trial (SCUT), a NIH-funded, multi-center clinical trial studying the effect of topical corticosteroids on bacterial corneal ulcers; 2) to obtain a Bayesian prior estimate for the effect of corticosteroids on clinical outcomes such as visual acuity in bacterial corneal ulcers.
Methods: Five hundred patients with culture-positive bacterial corneal ulcers were included in a prospective, double-masked trial of moxifloxacin plus either prednisolone phosphate 1% or placebo drops.  The primary outcome was best spectacle-corrected visual acuity at 3 months. Scenarios will be presented to the OMIG audience to calculate a prior estimate of the effectiveness of steroids in corneal ulcer treatment.
Results: Of the 485 patients in India and 15 in the US, 366 patients (73%) grew gram-positive and 140 (28%) grew gram-negative bacteria on cornea culture.  The most common organisms were S. pneumoniae (250, 50%), P. aeruginosa (111, 22%), and Nocardia spp (56, 11%). The most common in India were S. pneumoniae (250, 52%), P. aeruginosa (106, 22%), and Nocardia spp (56, 11.5%), and in the US, P. aeruginosa (5, 33%) and Moraxella spp (4, 27%).  Median baseline BSCVA was 20/125 for patients in India and 20/80 for patients in the US. Indian patients were most often employed as agricultural laborers (44%).  Contact lens use was reported in 3 patients in India (0.6%) and 5 patients in the US (33%).  Median infiltrate size was 2.65 mm for patients in India and 1.65 mm for patients in the US.
Conclusions:Patients in India were more likely to have worse visual acuity, larger ulcers, be infected with S. pneumoniae and Nocardia spp, and less likely to wear contacts.  Expert opinion solicited from OMIG attendees during the presentation will become the prior data for a secondary Bayesian analysis conducted post unmasking.

Disclosure Code: N; NIH funding: U10-EY015114-01