The aim of this study was to obtain the relevant information on antibiotic use in a 750-bed Croatian university hospital. The study has been designed as a 2-point prevalence interventional analysis. For each patient on antibiotic therapy, diagnosis, indication for treatment, antibiotic therapy, dosage and route of administration together with the results of microbiological studies (if available) were obtained. After the first prevalence analysis in 2001, a restriction on unlimited antibiotic use was introduced. The second analysis, performed in 2002, after restrictions on antibiotic use, revealed reductions in the rates of restricted release antibiotics and overall antibiotic use with decreases from 38.6% to 36.9% and 23.4% to 23.2% respectively (p = 0.87). The first survey showed that the 5 most often prescribed antibiotics in the therapy of bacterial infections were: gentamicin, other aminoglycosides, carbapenems, amoxycillin +clavulanate and vancomycin with proportions of 14.8%, 10.3%, 8.2%, 7% and 7% respectively. In the year 2002, the most prescribed antimicrobial drugs in the therapy of bacterial infections were: gentamicin, quinolones, vancomycin, carbapenems and cefuroxime with proportions of 18.6%, 11.4%, 9.7%, 9.3% and 8% respectively. A reduction in the proportions of doubtful antibiotic therapy, from 24.6% before the intervention, to 24.2% after the restrictions, accompanied by a 0.4% rise in the rates of indicated antibiotic therapy was also observed (p = 0.93). Our study shows that restrictions on formerly unlimited use of antimicrobials, even when leading to an improvement in their prescribing, do not necessarily cause rapid and significant reduction in the overall use of antibiotics or explicit positive financial effects.