Background The recent classification criteria sets for primary Sjögren's syndrome (AECG 2002 and ACR/EULAR 2017) did not include major salivary glands ultrasononography (SGUS).
Objectives The UTOPIA study was undertaken to determine if and how SGUS may improve the ACR/EULAR criteria.
Methods Twenty four international experts in pSS evaluated on an internet-secure relational database 512 randomly realistic vignettes derived from 150 patients with suspected pSS included in the french DiapSS cohort. Each vignette contained sections on “history” (duration of the symptoms, gender, age), clinical symptoms (dry mouth, dry eyes and systemic manifestations), results of the SGUS evaluation (score > ou < to 2), and results of the major tests to diagnose pSS (Schirmer's test, ocular staining score (OSS), salivary flow, focus score on salivary biopsy, presence of anti-SSA antibodies). Each expert had to score the diagnosis of pSS as absent, unlikely, likely or present for 64 vignettes. Each vignette was evaluated by 3 experts. Diagnosis of pSS was obtained when at least 2 of 3 considered it as likely or present. Univariate and multivariate analysis (Wald test) were performed to evaluate the association between the SGUS criteria, the ACR/EULAR criteria and its different individual items with the diagnosis of pSS as defined by the experts. Data were then replicated on independent cohorts of suspicion of pSS.
Results Univariate and multivariate analyses confirmed that ACR/EULAR criteria and SGUS were independently associated with the diagnosis of pSS. Disease duration, OSS and ocular dryness were not associated with the diagnosis of pSS. Only 6 variables were selected by logistic regression analysis: presence of anti-SSA (weight:4), focus score (weight:3), SGUS (weight:2), Schirmer's test (weight:1), dry mouth (weight:1) and salivary flow rate (weight:1). According to ROC curve analysis, a score of ≥5 had 96% Se and 84% Sp, compared with 90% Se and 84% Sp for the ACR/EULAR criteria. The corrected C statistic (AUC) for the new weighted score was 0.98.
Conclusions Inclusion of the SGUS item in the ACR/EULAR criteria improves their diagnostic performance.
Disclosure of Interest S. Jousse-Joulin: None declared, F. Gatineau: None declared, C. Baldini: None declared, S. Arends: None declared, F. Barone: None declared, A. Baer: None declared, H. Bootsma: None declared, S. Bowman: None declared, P. Brito-Zeron: None declared, D. Cornec: None declared, T. Dorner: None declared, S. De Vita: None declared, B. Fisher: None declared, D. Hammenfors: None declared, M. Jonsson: None declared, X. Mariette: None declared, V. Milic: None declared, T. Nakamura: None declared, W.-F. Ng: None declared, E. Nowak: None declared, A. Rasmussen: None declared, R. Seror: None declared, C. Shiboski: None declared, T. Nakamura: None declared, M. Ramos-Casals: None declared, A. Vissink: None declared, A. saraux: None declared, V. Devauchelle-Pensec Grant/research support from: ROCHE-CHUGAI