Comparison of three clinical scoring systems for Culicoides hypersensitivity in a herd of Icelandic horses

Culicoides hypersensitivity (CH), an intensely pruritic and seasonal allergic dermatitis, is a common allergic disease affecting horses worldwide. Currently, there is no validated clinical scoring system for the quantification of clinical signs associated with CH.

To (i) determine the best cut-off point of three scoring systems, (ii) test the accuracy of each system when compared to the clinical diagnosis of an experienced veterinarian and (iii) assess agreement between systems.

Icelandic horses (n = 20); eight with CH and 12 unaffected, from a research herd receiving no treatments for allergic dermatitis.

Lesion scores were recorded biweekly from April until September with three clinical scoring systems (A, B and C) by a single observer initially blinded to CH status. Separate logistic regression analyses for each time point were used to determine appropriate cut-offs for CH classification. Spearman's rho and Cohen's kappa were calculated to analyze correlation of scores and agreement of CH categorization between systems, respectively.

The best allergic cut-off scores for system A, B and C were determined to be three, eight and 12, respectively. For each system median areas under the curve (>0.85) were excellent and discriminatory ability for correctly classifying CH status was strong. Excellent correlation between scores for each system (Spearman's rho > 0.96) and excellent intersystem agreement for CH categorization (kappa ≥ 0.73) were found across scoring time points.

Results support the use of these scoring systems as templates for the future standardization of a CH clinical scoring system.