OBJECTIVE To describe the clinical picture of arthritis in patients with chronic infection by hepatitis C virus (HCV). METHODS Two patient populations were studied. Patients with arthritis and evidence of serum elevation of alanine aminotransferase (ALT) at the consultation were checked for HCV infection. A second group of 303 consecutive patients with rheumatoid arthritis (RA) were also checked for the presence of HCV antibodies. All patients attended the outpatient rheumatology unit of a tertiary care teaching hospital. Chronic HCV infection was determined by the presence of viral RNA in serum. A group of 315 first-time blood donors served as controls. RESULTS Twenty-eight patients with arthritis and chronic HCV infection were identified. Seven fulfilled criteria for RA, psoriatic arthritis was found in one patient, systemic lupus erythematosus in one, gout in 2, chondrocalcinosis in 2, osteoarthritis in 7, and tenosynovitis in one. In 7 patients with a clinical picture of intermittent arthritis, a definitive diagnosis could not be made. In these patients, mixed cryoglobulinemia was present in 6/7 (86%), whereas mixed cryoglobulinemia was found in 6/21 (28%) of the other patients. Among patients with RA, 23 (7.6%) had HCV antibodies, and active infection by HCV was found in 7 (2.3%) patients. The prevalence of HCV antibodies in a blood donor population was 0.95%, significantly different (p<0.001; 95% CI 0.03, 0.10) compared to patients with RA. The distribution of antibodies determined by recombinant immunoblot analysis was similar (p = NS) between RA patients and blood donors with HCV antibodies. CONCLUSIONS There is not a single clinical picture of arthritis in patients with chronic HCV infection. There is a well defined picture of arthritis associated with the presence of mixed cryoglobulinemia that consists of an intermittent, mono or oligoarticular, nondestructive arthritis affecting large and medium size joints. Although a high prevalence of HCV antibodies is suspected in patients with RA, its occurrence may be coincidental and its interpretation is difficult to determine from the data in this study.