Interobserver Agreement in Grading Activity and Site of Inflammation in Eyes of Patients with Uveitis

• Published in: American journal of ophthalmology 2008-12, Vol.146 (6), p.813-818.e1
• Main Authors: Kempen, John H, Ganesh, Sudha K, Sangwan, Virender S, Rathinam, Sivakumar R
• Format: Article
• Language: English
• Subjects: Agreements; Anterior Chamber - pathology; Biological and medical sciences; Clinical trials; Colleges & universities; Confidence intervals; Cross-Sectional Studies; Epidemiology; Eye diseases; Eyes & eyesight; Humans; Inflammation - classification; Medical sciences; Miscellaneous; Observer Variation; Ophthalmology; Reproducibility of Results; Uvea diseases; Uveitis - classification; Vitreous Body - pathology
• ISSN: 0002-9394; 1879-1891
• DOI: 10.1016/j.ajo.2008.06.004

Purpose To evaluate the reproducibility of new criteria for grading the site and activity of intraocular inflammation. Design Cross-sectional agreement study. Methods Grading of 202 eyes of 101 patients with uveitis was conducted by pairs of uveitis subspecialists at three uveitis subspecialty clinics. Agreement in grading location of inflammation, anterior chamber (AC) cells, AC flare, vitreous cells (present or absent), and vitreous haze was calculated. Results Proposed criteria for grading the location of intraocular inflammation had moderate reproducibility (κ range, 0.49 to 0.61). Reproducibility improved (κ range, 0.61 to 0.73) when the newly proposed category of anterior and intermediate uveitis was excluded. The ranges of κ statistics for exact agreement on gradings of AC cells (range, 0.34 to 0.43) demonstrated low to moderate levels of agreement, and gradings of AC flare (range, 0.50 to 0.64), vitreous cells (range, 0.48 to 0.51), and vitreous haze (0.53) were in the moderate agreement range. However, agreement within 1 grade was outstanding for AC cells (κ range, 0.81 to 1.00) and vitreous haze (κ, 0.75). For AC flare, a distribution skewed toward low grades within 1 grade made κ statistics unstable. Conclusions Proposed methods for grading inflammatory activity have moderate reproducibility for exact agreement in most instances. However, agreement within 1 grade is excellent for grading of AC cells and vitreous haze. The method for grading the site of intraocular inflammation also produces moderate levels of agreement, and in our hands was improved by excluding both anterior and intermediate uveitis. Improved methods for grading AC flare and vitreous cells are needed.