| Objective:To identify risk factors at diagnosis that can serve as prognostic indicators of clinical outcome in cases of fungal keratitis.Design:Retrospective cohort study.Methods:Patients with culture-proven fungal keratitis examined from January2009to December2012were included. Patients achieving microbiological cure with medical therapy alone were defined as had favorable clinical outcome; those achieving microbiological cure with medical therapy combined with therapeutic penetrating keratoplasty (T-PKP) were defined as had moderate clinical outcome; those who got refractory complications, or had no light perception, or got amotio retinae, or enucleated after medical and/or T-PKP were defined as had poor clinical outcome. An initial univariate analysis using Chi-Square Test was performed to identify and select important risk factors for subsequent inclusion in the regression model. Multinomal Logistic Regression was used to select the main risk factors for clinical outcome.Results:A total of154patients were included. Patients were treated with oral itraconazole, or combined with0.15%amphotericin B and/or voriconazole eye drop when in need. Of the154patients,126(81.8%) had favorable clinical outcome,18 (11.7%) had moderate clinical outcome,10(6.5%) had poor clinical outcome. Demographics and medical history including sex (p=0.998), age (p=0.175), preexisting ocular diseases (p=0.504), systemic diseases (p=0.178), trauma (p=0.372), recently corticosteroid use (p=0.195) were not significantly correlated with clinical outcome, while presenting days of symptom (p=0.054) was significantly correlated with clinical outcome. Presenting features of fungal keratitis including size of infiltration (p<0.001), depth of infiltration (p<0.001), hypopyon (p<0.001), position of the ulcer (p=0.004) were significantly correlated with clinical outcome. Fungal genus (p=0.046) and in vitro susceptibility (p=0.002) were both significantly correlated with clinical outcome. Using Multinomal Logistic Regression, we found that infiltration larger than6mm (p=0.003), infiltrate to the deeper two thirds of cornea (p=0.011), presence of hypopyon (p=0.044), and susceptible to no more than one drug in vitro (p=0.002), infection with Fusarium.spp or Aspergillus.spp (p=0.059) were the main risk factors for clinical outcome in fungal keratitis.Conclusions:In patients with fungal keratitis treated with oral itraconazole, or combined with0.15%amphotericin B and/or voriconazole eye drop when in need, large size of infiltration, deep infiltration, presence of hypopyon, infection with Fusarium.spp or Aspergillus.spp, and resistance to those drugs in vitro were the main risk factors for clinical outcome. |
