Objective: The presence of anti-citrullinated protein antibodies (ACPAs) in primary Sjögren's syndrome (pSS) ranges from 3% to 9.9%; however, there is no agreement about their clinical significance. Our aim was to systematically review the literature regarding the association of arthritis and ACPAs in pSS and their role in the development of rheumatoid arthritis (RA).Method: A comprehensive search of MEDLINE, ISI Web of Knowledge, and Cochrane Library from inception until June 2016 was undertaken using the combination of two or three of the keywords: primary Sjögren's syndrome, Sjögren's syndrome, arthritis, synovitis, arthropathy, anti-cyclic citrullinated peptide antibodies, and anti-citrullinated protein antibody – ACPA. No language restriction was used. Studies were included if they: assessed the association of arthritis and ACPAs, had sufficient data to construct a two-by-two table, tested immunoglobulin G ACPA by any method, and included patients with pSS according to a validated set of classification criteria. We used a random effects model and evaluated the heterogeneity and publication bias.Results: Ten studies were included (involving 1322 patients). We found a pooled odds ratio of 4.42 (95% confidence interval 1.15–16.94, p = 0.03). The test for heterogeneity was I2 = 0.87. Publication bias was not observed. Based on data from three studies, 33 of 58 pSS patients with ACPAs (57%) developed RA compared with none of 598 pSS patients with negative ACPA (p < 0.000001).Conclusion: Patients with pSS disclosing ACPAs are prone to arthritis as part of the clinical spectrum of the disease, but are also at risk of developing RA.