BACKGROUND Urinary tract infections are associated with substantial morbidity and recurrent infections. Antibiotic therapy is generally initiated empirically because early treatment decreases the rate of morbidity resulting from UTI. Unfortunately, antibiotic resistance has become an increasingly pressing problem in many countries. In this study, the resistance patterns of urinary isolates to commonly used antimicrobials were determined in order to evaluate the options for empirical antibiotic therapy of UTI in out- and in- patients. METHODS A retrospective study was carried out on urine samples examined in this laboratory in 2002-2005. The isolates were divided into the following three groups: isolates from hospital inpatients, isolates from community outpatients, and isolates from catheterized patients. RESULTS Escherichia coli was the most common etiologic agent isolated, followed by Enterococcus faecalis and Klebsiella pneumoniae. Over the four-year period, a decrease in the isolation of Pseudomonas aeruginosa and a parallel increase in Candida spp. in hospitalized patients were observed. Against Gram-positive isolates, enterococci in particular, ampicillin and glycopeptides demonstrated the best, most consistent activity. Among Escherichia coli isolates, nitrofurantoin, cephalosporins, and penicillin/betalactams were the best options for therapeutic treatment because of the presence of a rate of resistance to cotrimoxazole and fluoroquinolones of over 10%, while the most active drug against Pseudomonas aeruginosa was piperacillin/tazobactam. CONCLUSIONS Region-specific surveillance studies provide additional information about the type of pathogens causing UTIs and their antimicrobial susceptibility patterns. Therefore, these data can serve as a basis to develop national country-specific guidelines for the empirical treatment of UTIs.