Background Many classification criteria sets have been proposed for primary Sjögren's Syndrome (pSS),consisting of variable features of the disease.Despite increasing number of criteria sets, expert opinion is still keeping its importance for the diagnosis of pSS.
Objectives We aimed to compare the performance of 3 classification criteria sets for pSS, to determine the agreement between each other and expert opinion and to investigate diagnostic contribution with any combined use of these criteria.For those assessments, American-European Consensus Criteria (AECG), American Collage of Rheum (ACR)/Sjögren Int. Clinical Alliance (SICCA) criteria and L.S. Martin et al. criteria named Mathematical model (M.model) had chosen.The M.model is a criteria set known to be noninvasive and proposed as highly specific.
Methods 86 patients (F: 96%, mean age: 51.7±11.8) following up in our clinic with the diagnosis of pSS were enrolled to the study.Expert opinion had been taken as gold standard for pSS diagnosis. Patients were questioned for eye/mouth dryness. Antinuclear antibody (ANA), complement (C3, C4), anti-Ro and anti-La, serum protein electrophoresis, rheumatoid factor (RF), break-up time (BUT) and Schirmer, minor salivary gland biopsy results were analysed cautiously. The aforementioned criteria sets, were implemented to assess the classification. The percentage of exact agreement and Kappa test was calculated.
Results Number of patients classified as pSS according to ACR/SICCA, AECG and M.model criteria were 75 (87%), 63 (73%) and 58 (67%) respectively. 9 patients (10%) did not fulfill any of these 3 criteria but they were diagnosed as pSS according to expert opinion (Image 1). The Kappa test was moderate between the ACR/SICCA and AECG criteria and slightly low between the M.model with the AECG and ACR/SICCA criteria (Table 1). According to our proposed combination model, number of patients classified as pSS, either AECG or M.model was 74(86%); either AECG or ACR/SICCA model was 76(88%); either ACR/SICCA or M.model was 77 (89.5%) (Table 2). In the triple combination of our model, if all 3 criteria used concomitantly, only 46 (51%) of patients fulfill the criteria simultaneously; however if any of 3 criteria sets used 77 (89.5%) patients fulfill as pSS.
Conclusions In this study, ACR/SICCA classification criteria for pSS was found the most compatible criteria set with expert opinion, when it used alone. The combination of either ACR/SICCA or M.model was found to be the most sensitive binary combination for classification with similar results as triple combination.This study showed that a small group of patients could not be classified as pSS even if the criteria sets were used with concomitant combinations. It was determined that experienced specialists were largely need to recognize and distinguish these patients.
Disclosure of Interest None declared