Susceptibility Testing and Clinical Outcome in Fungal Keratitis

Abstract

Purpose: : Treatment for fungal keratitis remains largely empirical. Although techniques for determining fungal susceptibilities are improving, the correlation between susceptibility testing and clinical outcomes for corneal ulcers remains unknown. Fungal keratitis is notoriously difficult to manage and any test that facilitated tailoring of treatment to particular organisms would be valuable. Here we assess whether fungal susceptibility testing correlates with clinical outcome in cases of fungal keratitis.

Methods: : This is a retrospective case review of 98 consecutive patients with culture proven fungal keratitis presenting to the Aravind Eye Hospital between March and July 2004. Patients were treated with natamycin, and cultures from patients' corneas underwent macrobroth minimum inhibitory concentration (MIC) testing to natamycin. A healing time of <3 weeks from presentation was considered a good result. Multiple logistic regression was used to predict a good/poor result.

Results: : The MIC, organism genus (Fusarium versus Aspergillus), and infiltrate size were correlated to the primary outcome of ulcer healing at 3 weeks. Lower MIC was significantly associated with a good outcome as were Fusarium species and a smaller presenting infiltrate. Lower MIC significantly predicted incidence of perforation (OR = 2.31, P = 0.01). However, in a multivariate, logistic regression model including infiltrate size at presentation, MIC, and organism as covariates, only ulcer size at presentation appears to correlate with outcome (OR = 0.31, P = 0.004).

Conclusions: : Antifungal susceptibility testing has been shown to predict outcomes in systemic diseases such as mucosal candidiasis and candidemia_, and antifungal susceptibilities have begun to influence treatment recommendations. In this study, a two-fold increase in MIC was associated with a 47% reduction in the odds of healing. Susceptibility, type of organism and size of infiltrate may all contribute to the likelihood of treatment success, but our relatively small study lacked sufficient power to assess the relative contribution of multiple predictors. In conclusion, we provide further evidence that susceptibility testing in fungal keratitis, like bacterial keratitis, may have a role in treatment algorithms although further study is necessary.

Keywords: keratitis • fungal disease • wound healing