Distinguishing Features of Ocular Sarcoidosis in an International Cohort of Uveitis Patients

To determine which clinical features distinguish ocular sarcoidosis from other forms of uveitis in an international population and to estimate the sensitivity and specificity of the International Workshop on Ocular Sarcoidosis (IWOS) clinical signs and laboratory tests. Multicenter, retrospective medical record review. Eight hundred eighty-four patients with uveitis from 19 centers in 12 countries. Data collected included suspected cause of uveitis, clinical findings, and laboratory investigations within 6 months of presentation. The IWOS criteria were used to classify patients as having definite (biopsy-proven), presumed (evidence of bilateral hilar lymphadenopathy [BHL] on chest radiograph or CT scan), probable, or possible ocular sarcoidosis. Patients with biopsy positive results or BHL on chest radiograph or CT scan were considered sarcoidosis cases. Sensitivity and specificity of clinical signs and laboratory investigations for diagnosing ocular sarcoidosis. Of the 884 uveitis patients, 264 (30%) were suspected to have ocular sarcoidosis. One hundred eighty patients (20%) met the IWOS criteria; 98 were definite (biopsy-proven) disease, 69 presumed disease (BHL), 10 probable disease, and 3 possible disease. Among sarcoidosis cases, the most common clinical signs were bilaterality (86%); snowballs or string of pearls (50%); mutton-fat keratic precipitates, iris nodules, or both (46%); and multiple chorioretinal peripheral lesions (45%). Sixty-two percent of sarcoidosis cases had elevated angiotensin converting enzyme or lysozyme and 5% demonstrated abnormal liver enzyme test results. Of the patients suspected of having sarcoidosis, 97 (37%) did not meet the IWOS criteria. With the exception of BHL, IWOS clinical findings and investigational tests had low sensitivities for diagnosing ocular sarcoidosis. In particular, liver function tests seem to have little usefulness in diagnosing ocular sarcoidosis. Many patients suspected of having sarcoidosis did not fit into the classification system, indicating that the guidelines may need to be reconsidered. Adding novel laboratory tests and using more advanced statistical methods may lead to the development of a more generalizable classification system.