The aim of this study was to determine prevalence of prescribing potentially inappropriate medications (PIMs) in elderly patients before and after hospitalization. Beers and STOPP criteria were used to detect potentially inappropriate medications. A cross-sectional observational study was conducted in University Hospital Dubrava. We included acutelly and electively admitted inpatients aged 65 years and older who were taking one or more drugs and who had signed consent to participate in research. PIMs were determined using two different protocols: Beer’s criteria and STOPP criteria. The study included 180 patients with a median age 75,0 years (range 65-96). On average, the patients spent 6,1 medications (range 1-19) and had a diagnosis of 7,8 (range 0-15). The Beers criteria detected 352(24,9%) of PIMs before hospitalization that were prescribed to 150 (83,3%) patients while after the hospitalization there were slightly fewer medications, in total 310(21,9%) that were prescribed to 133 (73,9%) patients. The STOPP criteria detected 175 (12,4%) of PIMs before hospitalization that were prescribed to 107 (59,4%) patients while after the hospitalization there was 155 (10,9%) of PIMs that were prescribed to 102 (56,7%) patients. A Beers criteria proved to be better tool in the detection of PIMs.The total number of PNLs before the hospitalziation is higher than the number of PNLs at release (Beers: 352 vs 310; STOPP: 175 vs 155) According to both protocols, the most prevalent inappropriate drug was diazepam, and the rest of them were acetylsalicylic acid, pantprazole, furosemide, ibuprofen and bisoprolol.Using different protocols in detecting potentially inappropriate medications causes more rational prescribing, improves life quality and reduces the overall cost of treatment.