OBJECTIVES: To evaluate the share of antibiotic therapy cost in total treatment costs for nosocomial and community acquired infections. METHODS: Systematic literature search (1996-2003) using the major online databases and additional manual search yielded 1211 references (selection Ia). Studies were divided up to 11 different diagnostic groups that were considered to be most relevant (selection Ib). Inclusion and exclusion criteria were applied for the selection Ic. Study quality was assessed with a consolidated quality score comprising 23 questions. A study was assessed as 'qualified‘ for further analyses, if at least 50% of maximum points were achieved in a review conducted by two independent reviewers. This resulted in the selection of 44 studies (selection II) that were subject to a detailed metaanalysis. RESULTS: The percentage of antibiotic costs in relation to total direct costs was low: nosocomial pneumonia (N=3, 1.4-13.7%, SD: 5.5), respiratory tract infections (N=2, 10-24.8%, SD: 5.3), community acquired pneumonia (N=14, 0.64-57.84%, SD: 24.65), chronic bronchitis (N=6, 1.1- 66.66%, SD: 17.19), urinary tract infection (N=3,5-33.82%, SD: 14.85), intraabdominal infection (N=2, 3.2-19.38%, SD: 5.26), surgical prophylaxis (N=2, 0.1-83.8%, SD: 33,02), otolaryngological infections (N=4, 14.18-60.1%), STD (N=1, 0.15-3.3%, SD: 2.23), Helicobacter pylori infections (N=3, 2.34-45.49%, SD: 16.36), Clostridium difficile infections (N=3, 0.07-1.20%, SD: 0.51). CONCLUSIONS: This review showed that costs of antibiotic treatment were low compared to overall direct treatment costs for most frequent infections regardless of the differences in design and quality of studies. The experience from this review may also contribute to further development of evidence-based guidelines for conducting pharmacoeconomical studies.