Abstract
Purpose: :
To determine if fungal and bacterial ulcers can be distinguished using only clinical features.
Methods: :
A group of 80 photographs representing corneal ulcers with laboratory confirmation of etiology were selected from the Steroids for Corneal Ulcers Treatment (SCUT) and Mycotic Ulcer Treatment Trial (MUTT) database. Forty bacterial ulcers and 40 fungal ulcers were randomly selected for inclusion. Along with the photographs, a survey was presented that asked clinicians to predict culture results, gram stain results, and to rank traits which were most helpful to them in making their decision. Two groups of clinicians were asked to evaluate the photographs and complete the survey: 8 clinicians from the Proctor Foundation cornea faculty and 7 clinicians from the Aravind Eye Hospital cornea service.
Results: :
79 of 80 photographs were included for analysis, one photograph was excluded as it was a mixed infection. Overall, clinicians were able to accurately distinguish bacterial from fungal etiology 65.6% of the time (95% CI, 63.3 to 67.9). The gram stain, genus, and species were accurately predicted 48.2% (95% CI, 40.9 to 52.0), 25.7% (95% CI, 21.2 to 28.7), and 20.9% (95% CI, 14.3 to 25.1) of the time, respectively. The ability of the two groups to distinguish bacterial from fungal ulcers were better than chance, with the Aravind group successful 68.2% (95% CI, 65.0 to 71.3, P<0.0001 compared to chance) of the time versus 63.5% (95% CI, 59.9 to 67.0, P<0.0001) with the Proctor group. Clinicians in India performed better (P=0.03) The traits most commonly used by clinicians to differentiate ulcers, irrespective of correct identification of etiology, were infiltrate border appearance, surrounding stromal haze, and presence of hypopyon.
Conclusions: :
While corneal specialists perform better than chance, they correctly differentiated bacterial from fungal corneal ulcers less than 70% of the time. Clinicians in an area with a high incidence of fungal keratitis performed marginally better. As expected, more specific categorization led to less successful clinical distinction. This study highlights the importance of obtaining appropriate microbiological testing during the initial clinical encounter.
Keywords: cornea: clinical science • keratitis • differentiation