Antibiotic use was evaluated in two cohorts of intensive care units (ICUs) in Germany. One cohort included ICUs participating in a surveillance programme (N=34) collecting antibiotic use and bacterial resistance data, with quarterly feedback. The second ICU cohort was from a cross-sectional study and represented a sample from hospitals in South-west Germany (N=58). Two dose definitions were used. These were the World Health Organization/Anatomical Therapeutic Chemical Classification (ATC) 2001 definition of defined daily dose (DDD), and a definition of recommended daily dose (RDD) that better reflected the currently prescribed dosages of parenteral drugs for hospitalized patients. Data were expressed as DDD or RDD per 100 patient-days. It was determined whether hospital size and affiliation, year of study and ICU type had an influence on overall use of antibiotics. Overall use differed between the two ICU cohorts irrespective of the dose definitions used. High use of antibiotics was primarily associated with hospital affiliation (university vs non-university) and hospital size. Mean overall use of antibiotics in non-university hospital ICUs ranged between 106 and 111 DDD/100 (59 and 67 RDD/100) for different hospital size categories, compared with 140 DDD/100 (87 RDD/100) in university hospital ICUs. In conclusion, in order to compare the use of antibiotics between ICU cohorts and to assess trends over time, data adjustment is required for hospital affiliation and size.