In 1991, updated recommendations for the administration of antimicrobial prophylaxis in surgery were made extensively available to all surgical wards of our hospital. Two years later we surveyed the implementation of these recommendations in our institution. Inpatients undergoing surgical procedures during a 6 month period were prospectively evaluated for the indication, type, timing and duration of antibiotic prophylaxis. On each day of the study, data of patients who underwent surgery in the previous 24-48 h were obtained and monitoring was continued for up to 5 days. Of the 215 evaluated patients 193 (90%) received prophylaxis. This rate was similar for elective and emergency surgeries. The rate of prophylaxis in surgeries for which its use is recommended was significantly higher than in surgeries where compelling data for its use are not available (96 vs. 74%, P = 0.000006). However, many and important deviations of basic principles of antimicrobial prophylaxis in surgery were found. Prophylaxis was administered systematically in some types of surgery lacking compelling data for its use. In almost 50% of the surgical procedures the first dose of antimicrobial prophylaxis was not administered at the optimal timing; prophylaxis was continued for more than 24 h in 21% of the cases, and the use of unstandardized regimens was common. Despite the availability of local guidelines, the implementation of a hospital program with regard to antibiotic prophylaxis in surgery may be difficult. Further measures should be applied to achieve this goal.