Effect of pretreatment with antifungals on clinical outcomes in the Mycotic Ulcer Treatment Trial I

Abstract

Purpose: To determine if pretreatment with antifungal medications is predictive of worse clinical outcome in a fungal keratitis clinical trial.

Methods: The Mycotic Ulcer Treatment Trial I was a randomized, double-masked, NEI-funded trial to determine the optimal treatment for filamentous fungal keratitis. 323 smear-positive fungal ulcer cases with enrollment visual acuity of 20/40 (0.3 logMAR) to 20/400 (1.3 logMAR) were randomized to receive 5% topical natamycin or 1% topical voriconazole at the Aravind Eye Care System in India. Prior topical and systemic antifungal use, dose and duration were collected at enrollment. Corneal scrapings were obtained for smears and cultures at enrollment and day 7 of treatment. Enrolled patients had positive smears for filamentous fungus and negative Gram stain for bacteria. Pre-specified outcomes included 3-month visual acuity (primary), 3-month infiltrate or scar size, corneal perforation and/or transplant, and re-epithelialization time.

Results: Of the 323 patients enrolled, 44% presented on an antifungal with 24% using only topical natamycin, 5% using only topical or systemic azole, and 15% using both topical natamycin and an azole. At enrollment, patients pretreated had on average 9 days of symptoms compared to 6 days for those not pretreated (P<0.001). Pretreated patients had larger mean baseline infiltrate size (P<0.001) and larger mean baseline epithelial defect size (P=0.02) compared to those not pretreated. Pretreated ulcers were significantly less likely to be fungal-culture positive at day 1 (P=0.04) and day 7 (P=0.046) compared to ulcers not pretreated. Multivariate regression analysis demonstrated that pretreatment with an antifungal was associated with significantly worse 3-month visual acuity (0.36 logMAR, 95% CI 0.11 to 0.62, P=0.006), larger 3-month scar size (1.47mm, 95% CI 0.82 to 2.13, P<0.001) and an increased odds of perforation and/or corneal transplant (OR 12.87, 95% CI 2.68 to 61.70, P=0.001). Pretreatment was not significantly associated with longer time to re-epithelialization.

Conclusions: In our analysis, we found that patients who were pretreated had longer duration of symptoms, had worse clinical characteristics at presentation, and were more likely to be fungal-culture negative. Study participants who were pretreated had significantly worse clinical outcomes, likely as a result of initial ulcer severity.