Background Salivary gland dysfunction is one of the most common features of Sjögren' syndrome (SS). Sialoscintigraphy (sSC) is the only technique nowadays available to objectify salivary gland dysfunction that provides simultaneous information about type of functional defect, distribution and severity. In recent years its interpretation based mainly on qualitative indices and the appearance of the American College of Rheumatology (ACR) criteria have shifted in favor of the biopsy (Bp). Some studies have analyzed the role of quantitative scintigraphic indices, mainly excretion fraction (EF%), in the diagnosis of SS, with mixed results.

Objectives 1) To describe glandular involvement patterns and EF% values useful in SS diagnosis. 2) To compare diagnostic accuracy of scintigraphic indices and Bp.

Methods Patients with suspected SS and with sSC and Bp performed between 2006–2012 were included. sSC were performed in a 30 minutes protocol with 99mTc sodium pertechnetate and lemon juice stimulation. Patients already diagnosed of SS and/or with sSC or Bp requested by specialists other than rheumatologists were excluded. Correlations between glandular involvement patterns and EF% and diagnosis of SS were analyzed with Fisher exact test and Kruskal Wallis respectively. Sensitivity, specificity and receiver operating characteristic curve analysis of scintigraphic indices and Bp were performed.

Results Among 272 patients with sSC performed in that period, 71 fulfilled inclusion criteria. 68 (96%) were women with a mean age of 52 ± 14 in the timing of sSC. 57 (80%) reported xerostomia, the majority (39%) with onset in recent months. Qualitative sSC were reported as abnormal in 51 cases (71%); data on age, xerostomia duration, xerostomia related drugs and presence of thyroid disease showed similar frequencies among normal and abnormal sSC groups. Sufficient data to apply American-European Consensus Group and ACR criteria were available in 69 and 48 patients respectively, with diagnosis of SS in 38 (54%). Uptake dysfunction (96 vs 77% p 0.134) and submandibular involvement (59 vs 30% p 0.236) were observed most frecuently in SS patients. All glands showed lower EF% medians in this group, with significant differences for submandibulary ones: right gland 29 (7–37) vs 49 (41–57) p 0.004; left gland 30 (10–45) vs 52 (29–64) p 0.007. Submandibular EF% (right gland AUC 0.793 ± 0.091 p 0.004 IC 95% [0.61–0.97]; left gland AUC 0.764 ± 0.083 p 0.007 IC 95% [0.60–0.93]) showed better diagnostic accuracy than qualitative sSC (AUC 0.601 ± 0.070 p 0.149 IC 95% [0.47–0.74]) and similar to Bp (AUC 0.789 ± 0.055 p 0.000 IC 95% [0.68–0.90]). Submandibulary EF% cuttoff values of 39.5 (right gland) and 47.5 (left gland) showed a sensitivity of 95–82% and specificity of 71–67%, compared to 82% and 39% for qualitative sSC, and 58% and 100% for Bp.

Conclusions 1) Patients fulfilling SS criteria show lower submandibular EF% compared to those who do not. 2) Diagnostic accuracy of submandibular EF% is better than qualitative sSC and similar to Bp. 3) Cutoff values of submandibular EF% of 39.5 (right gland) and 47.5 (left gland) show a sensitivity comparable to qualitative sSC but a significantly higher specificity.

Disclosure of Interest None declared