Reproducibility of the classification of ocular ferning patterns in Sjogren's syndrome patients

Sergio Felberg Hamilcar Cordeiro Elcio Hideo Sato Dino Martini Filho Maria Cristina NIshiwaki-Dantas Roberto Mitiaki Endo Paulo Elias Correa Dantas About the authors

PURPOSE: To verify the reproducibility of Rolando's classification of the tear ferning test using five different examiners and to compare the patterns of crystallization found in Sjögren's syndrome patients and normal subjects. METHODS: Tear ferning analysis of 29 patients with Sjögren's syndrome and of 45 patients without ocular disease were done using polarized light microscopy and the Rolando classification for tear ferning. Five examiners classified the ferning patterns of all the patients. ROC curve (Receiver Operating Characteristic) was used to find out the best score for the correct syndrome diagnosis. Kappa index (p<0.0001) was used to compare the results of the examiners among them and check the test's reproducibility. Charts were drawn to compare the two groups' results. RESULTS: Throught the ROC curve the score of 2.50 for diagnosis of Sjögren's syndrome was stabilished. Considering the aggregated patterns I with II and III with IV, the examinors' level of pattern agreement was excellent (Kappa ranging from 0.82 to 0.97, p<0.0001). The group with Sjögren's syndrome was classified mostly as patterns III and IV and the patients without ocular disease mostly as I and II. CONCLUSION: The patterns associated with Sjögren's syndrome and normal patients matched the ones in the literature. The tear ferning test classification is reproductible when the Rolando classification was used for Sjögren's syndrome patients.

Sjögren's syndrome; Crystallization; Crystallization; Tears; Diagnostic techniques, ophthalmological; Reproducibility of results

Conselho Brasileiro de Oftalmologia Rua Casa do Ator, 1117 - cj.21, 04546-004 São Paulo SP Brazil, Tel: 55 11 - 3266-4000, Fax: 55 11- 3171-0953 - São Paulo - SP - Brazil