The purpose of this study was to investigate the influence of an antibiotic policy programme based on restriction of the empirical use of fluoroquinolones and ceftazidime on the susceptibilities of Gram-negative microorganisms in a general Intensive Care Unit (ICU). The epidemiology of infections caused by the predominant ICU pathogens, i.e. Acinetobacter baumannii, Pseudomonas aeruginosa and Klebsiella pneumoniae, and their resistance patterns as well as antibiotic consumption were recorded for a 6-month period. An antibiotic restriction policy including ceftazidime and quinolones was applied. After an 18-month period of protocol application, the same parameters were recorded for another 6-month period. Consumption of restricted and overall antibiotics was reduced by 92.5% and 55.4%, respectively. Susceptibilities to ciprofloxacin of the three predominant infection-causing Gram-negative bacilli were significantly increased. Ceftazidime showed an increase in susceptibility only for P. aeruginosa. Similar rates of infectious episodes were recorded in the two periods and no differences were observed either in overall mortality or in ICU ecology as expressed by the type of microorganisms implicated in colonisation and/or infection. The reported data suggest that an antibiotic restriction policy can significantly reduce antimicrobial consumption and antimicrobial resistance rates, although the latter effect can be also influenced by the prevalent resistance mechanisms and the prevalence of imported resistance.