ABSTRACT
Purpose: To analyze the diagnostic predictive ability of clinical variables.
Methods: Demographic and clinical variables of 172 serologically proven leptospiral uveitis patients were compared with 200 controls of non-leptospiral uveitis. Multiple logistic regression analysis identified diagnostic predictors. A receiver operating characteristic curve tested the performance of the model.
Results: Of all variables, male gender, farming as an occupation, and clinical features such as non-granulomatous panuveitis, hypopyon, and vitreous infiltration in the absence of retinochoroiditis constituted the predictive parameters, with the sensitivity and specificity of 86% and 90.7%, respectively.
Conclusions: Multiple logistic analysis detected clinically diagnostic predictors that can assist primary care ophthalmologists. Clinical diagnosis can further be confirmed by serology at tertiary care centers.
DECLARATION OF INTEREST
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.