This multicenter retrospective study explored clinical features that distinguish ocular sarcoidosis from other forms of uveitis. The authors also estimated the sensitivity and specificity of the International Workshop on Ocular Sarcoidosis (IWOS) guidelines for characterizing uveitis patients suspected of having sarcoidosis.
Study design
Medical records from 884 uveitis patients from 19 centers in 12 countries were included in the final analysis. Of these, 167 had sarcoidosis that was confirmed from a positive biopsy or from evidence of bilateral hilar lymphadenopathy (BHL) on chest radiograph or CT scan.
Outcomes
Among sarcoidosis cases, the most common clinical findings were bilateral disease (86%), snowballs or string of pearls (50%), mutton-fat keratic precipitates, iris nodules or both (46%) and chorioretinal peripheral lesions (37%). As expected, chest x-ray, angiotensin converting enzyme (ACE) levels and lysozyme levels were neither sensitive nor specific.
The authors compared their findings to the IWOS diagnostic criteria and found that 97 of 167 (37%) patients suspected of having sarcoidosis did not meet the IWOS criteria.
Limitations
Although this study was retrospective, the authors attempted to overcome this limitation with a large uveitis control group. The study was also inadequately powered to evaluate differences between populations.
Clinical significance
This study reminds us not to rely on chest x-ray and lysozyme and ACE levels to rule out sarcoidosis. Even CT of the chest is not always positive in biopsy-proven sarcoidosis.
It is important to note that the IWOS criteria were strongly influenced by features described in East Asian patients, so it is unsurprising that the IWOS was somewhat lacking in this study with a diverse cohort. The findings here remind us that the diagnosing ocular sarcoidosis depends on a cluster of findings. Despite some drawbacks, the IWOS guidelines do recognize the need to assess a variety of clinical signs and laboratory findings. Diagnostic criteria are notoriously difficult to devise, and are often designed for consistency in reporting research than necessarily applying to the patient in your chair. With no definitive diagnostic test for ocular sarcoidosis, we await more advanced techniques, such as latent class analysis or machine learning, which may lead to a more generalizable classification system.