Background/Purpose: Sjögren’s syndrome (SS) is a complex autoimmune disorder characterized by xerostomia and xerophthalmia due to exocrine gland dysfunction. There is no single diagnostic test for SS and multiple research classification criteria have been proposed. Currently, a combined EULAR-ACR working group is dedicated to resolve discrepancies and weaknesses between two systems presently in use: the American-European Consensus Group (AECG) criteria and the SICCA (ACR) criteria, with the ultimate goal of establishing a new consensus classification. A point of contention has been the assessment of keratoconjunctivitis sicca either by the van Bijsterveld score (vBS) for AECG classification or the Ocular Staining Score (OSS) for ACR classification. We present a direct comparison of the two scoring systems to help clarify the matter. 

Methods: We performed all tests for AECG and ACR classification in a multidisciplinary sicca clinic. Complete vBS and OSS evaluations were available for 716 participants; a subset of 587 were classified by AECG criteria either as pSS (n=257) or sicca (n=330).  The remaining 129 had other diseases or overlap/secondary SS. Initial analysis of concordance (vBS=OSS) or discordance (vBS≠OSS) of the ocular scores was done for n=716 subjects but the correlations with classification criteria and clinical features was restricted to the pSS/sicca subset. 

Results: Of the 716 subjects, 538 (75.1%) were concordant while 178 (24.9%) were discordant for vBS vs. OSS. The discordant subjects had significantly higher vBS (Wilcox rank sum p<2.2x10E-16); the same held true if only pSS vs. sicca were compared. ROC curves comparing the sensitivity and specificity of the vBS and OSS both in the two study groups showed that the accepted vBS cutoff of 4 has a sensitivity of 0.59-0.68 and specificity of 0.74-0.79; similar sensitivities for the OSS are observed at scores of 4 (sensitivity 0.63-0.74; specificity 0.72-0.78) and 5 (sensitivity 0.54-0.62; specificity 0.79-0.83). Discordant participants were significantly more Ro (+), La (+), and biopsy (+) than the concordant cases (p=1.7xE10-10; 4.3×10-6; 1.8x10E-10 respectively). The patches of confluent staining were the most common cause of discordance in the scores. When analyzing the three additional corneal staining points of the OSS, their presence was highly associated with participants meeting criteria for pSS (p=8.4x10E-7 to 1.73x10E-13); with (+) Schirmer’s (p=3.1x10E-6 – 6.5x10E-10), Ro (p=2.7x10E-5 – 2.9x10E-10), La (p=0.01 – 8.8x10E-6), biopsy (p=1.8x10E-5 – 3.4x10E-10), and WUSF (p=0.0007 – 1x10E-6). In all cases, the patches of confluent staining were the most highly associated with markers of disease severity while the corneal filaments were the least significant. 

Conclusion: The OSS was introduced as an objective measure of KSS in the ACR classification and was considered abnormal if ≥3. However, more recent studies indicate it has poor specificity. Our results in a large sicca cohort suggest that a cutoff between 4 and 5 would maintain a good sensitivity while increasing the specificity significantly; a matter of great importance when applying the criteria for patient selection for clinical trials.

---

« Back to 2014 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/performance-of-the-ocular-staining-score-oss-vs-the-van-bijsterveld-score-in-the-assessment-of-sjogrens-syndrome-related-keratoconjunctivitis-sicca/