Objective: Outpatient antibiotic use in Germany differs substantially between eastern and southern parts of the country (low use) and the western part (high use). Whether similar regional variation exists in hospital antibiotic consumption is not known. We investigated this issue using a convenience sample of 145 hospitals providing data for the year 2003. Methods: Data on hospital consumption of systemic antibiotics in Anatomical Therapeutic Chemical (ATC) class J01 were obtained from acute care hospitals that participated in an IMS survey and had complete data (dispensed drugs and patient-days per year) for at least one non-pediatric, non-psychiatric department or ward. A total of 275 non-ICU surgical departments/wards, 229 non-ICU non-surgical (general medicine, haematology-oncology, neurology/stroke) departments/wards, and 184 ICUs were analysed. Data were expressed in DDD (WHO/ATC definition version 2003) or daily doses adapted for recommendations in hospitalized patients (RDD) per 100 patient days (DDD/100 and RDD/100). Findings: The weighted mean over all departments/wards was 49.6 DDD/100 or 31.3 RDD/100, respectively. As expected, ICU antibiotic use density was much higher than use in non-ICU areas, and use in haematology-oncology was higher than in other non-surgical departments/wards. In univariate analyses, region, hospital bed-size category, university affiliation and haematology-oncology as specialty were associated with use density, but these associations were only partly confirmed in multivariate logistic regression analyses of factors associated with excess (≥ 75%) use density which showed university affiliation and haematology-oncology but not hospital location to be independently associated with comparatively high use. Conclusions: Antibiotic use density in German acute care hospitals does not appear to differ significantly between regions. Overall hospital consumption of antibiotics in this country appears to be similar to what has been described from other parts of Europe. In comparative analyses of hospital antibiotic consumption, data need to be adjusted at least for university affiliation and haematology-oncology.