Clinical differentiation of Pythium keratitis from fungal keratitis and development of a scoring system

PURPOSE: To differentiate Pythium keratitis from fungal keratitis using clinical signs, to explore usefulness of various signs as diagnostic prognosticators, and develop a clinical scoring system. METHODS: A retrospective review of medical records and archived clinical photographs of patients with culture-positive Pythium keratitis and hyaline filamentous fungal keratitis was conducted at a tertiary eye institute to explore characteristics of ulcers that may aid diagnosis. RESULTS: Full-thickness corneal stromal keratitis (P = 0.055), a dry ulcer surface (P = 0.010), tentacles (P < 0.0001), intrastromal dots (P < 0.0001), ring infiltrates (P = 0.024), reticular patterns (P < 0.0001), and peripheral furrows (P < 0.0001) were clinical signs associated with Pythium keratitis. Multiple regression analysis identified tentacles (odds ratio: 24.1, 95% confidence interval (CI): 3.8–158.1, P = 0.001) and peripheral furrows (odds ratio: 60.6, 95% CI: 5.1–712.3, P = 0.001) as independent diagnostic prognosticators for Pythium keratitis. The positive and negative likelihood ratios of a dry ulcer surface, tentacles, intrastromal dots, ring infiltrates, reticular patterns, and peripheral furrows predicting Pythium keratitis were 1.6, 13.6, 17.9, 4.3, 30.7, 15.3 and 0.4, 0.4, 0.7, 0.9, 0.6 and 0.8, respectively. The presence of two or more of these clinical signs (excluding a dry ulcer surface) had a sensitivity of 55.6% and a false positive rate of 1.4%. CONCLUSION: Tentacles, intrastromal dots, ring infiltrates, reticular patterns, and peripheral furrows are clinical signs to be considered for the diagnosis of Pythium keratitis and the presence of two or more signs has a very low false positive rate.