OBJECTIVE To determine the prevalence of and identify drug-drug interactions (DDI) between non-steroidal anti-inflammatory drugs (NSAID) and other drugs in a prescription database of patients with rheumatic iseases. METHODS This is a cross-sectional study of a drug prescription database saving information on 35,000 individuals who benefited from a pre-paid medical system serving bank employees and their relatives. The analysis included one year period (from January to December 1998). NSAID-DDI were listed and classified into 3 levels (1: minor; 2: moderate, and 3: high health risk or death risk) according to DRUGDEX(r) as well as MEDLINE and EMBASE search. RESULTS We analyzed 3,207 NSAID prescriptions (1.7±1.6 per patient) to 1,855 rheumatic patients (adults: 76.7%; geriatric: 20.2%, and pediatric: 3.0%; soft tissue rheumatism: 52%; osteoarthritis: 19%, and rheumatoid arthritis: 10%). There were 648 (20.20%) NSAID-DDI prescriptions: 594 (91.66%) corresponded to level 1; 46 (7.09%) to level 2, and 8 (1.23%) to level 3. In addition, 96 (2.99%) prescriptions included NSAID duplications. Interestingly, we found no NSAID-DDI with anticoagulants, anticonvulsants, and oral hypoglycemiants were found. CONCLUSIONS The prevalence of NSAID-DDI prescriptions to 1,855 rheumatic patients was 20.20% in one year. NSAID-DDI was mostly (91.66%) level 1, and rarely (1.23%) level 3. NSAID duplications were found in 2.99%. These results provides information on the frequency of prescriptions with DDI, which might potentially produce harmful effects and data, which may help in the development of studies searching for the clinical relevance of NSAID-DDI.