This study proposes to measure frequency and to characterize the profile of potential drug interactions (pDDI) in a general medicine ward of a teaching hospital. Data about identification and clinical status of patients were extracted from medical records between March to August 2006. The occurrence of pDDI was analyzed using the database monographs Micromedex® DrugReax® System. From 5,336 prescriptions with two or more drugs, 3,097 (58.0%) contained pDDI. The frequency of major and well document pDDI was 26.5%. Among 647 patients, 432 (66.8%) were exposed to at least one pDDI and 283 (43.7%) to major pDDI. The multivariate analysis identified that factors related to higher rates of major pDDI were the same age (p< 0.0001), length of stay (p< 0.0001), prevalence of hypertension [OR=3.42 (p< 0.0001)] and diabetes mellitus [OR=2.1 (p< 0.0001)], cardiovascular diseases (p< 0.0001) and the number of prescribed drugs (Spearman’s correlation=0.640622, p< 0.0001). Between major pDDI, the main risk was hemorrhage (50.3%), the most frequent major pDDI involved combination of anticoagulants and antiplatelet drugs. Among moderate pDDI, 3,866 (90.8%) involved medicines for the treatment of chronic non-communicable diseases, mainly hypertension. In HU-USP, the profile of pDDI was similar among adults and elderly (the most frequent pDDI and major pDDI were same), the difference was only the frequency in either group. The efforts of the clinical pharmacists should be directed to elderly patients with cardiovascular compromise, mainly in use of anticoagulants and antiplatelet drugs. Furthermore, hospital managers should increase the integration between levels of health care to promote safety patient after discharge.