OBJECTIVE To investigate whether patients with rheumatoid arthritis (RA) have immunological or clinical evidence of gluten hypersensitivity. METHODS Antigliadin antibodies (AGA) and antireticulin antibodies (ARA) were determined in two groups of RA patients and in a control group of patients with spondylarthropathies. In the first group of 42 patients with recent-onset RA, AGA and ARA were studied longitudinally during a one-year follow up period. In the second group of 36 patients with advanced RA and various abdominal symptoms examined by upper gastrointestinal endoscopy, AGA and ARA were determined cross-sectionally. RESULTS Increased AGA (IgA or IgG) levels were found in 37% (29/78) of all RA patients compared to 12% (3/25) of controls. ARA positivity (IgG) was found in 4% (3/78) of RA patients and in none of the controls. AGA positivity was increased in patients with early RA compared to patients with advanced disease (48% vs. 25%) but the difference was not statistically significant. However, no true gluten hypersensitivity with positive AGA and ARA together with villous atrophy was observed. CONCLUSIONS Despite the increased AGA positivity found distinctively in patients with recent-onset RA, none of the RA patients showed clear evidence of coeliac disease. AGA positivity in early RA may indicate a role of the gut immune system in the initiation of RA.