Ba!ckground: Healthcare associated infections, commonly observed in the intensive care unit (ICU), are strongly correlated with increased mortality, morbidity and medical costs. To address this, the contribution of both prophylactic and therapeutic antibiotics is incontestable. However, in ICU, the empirical choice of antibiotics is in 50% cases inappropriate and its use has often an arbitrary duration. This engenders not only undesirable effects that can prove to be fatal; but eventually favors the emergence of multi-resistant bacteria, and even Pan-resistant bacteria. Therefore, the efficient use of antibiotics in terms of their initiation, duration and regimen re-evaluation must be correctly managed. In this study, we look at antibiotic therapy data, especially the motivating factors of changing antibiotic therapies in the peri-ICU stay period. Material and Methods: Retrospective this study included patients aged ! 16 years, who had an ICU stay ! 72 hours and had received antibiotics. Clinical data were collected and the following scores calculated: logistic organ dysfunction (LOD), sequential organ failure assessment (SOFA), McCabe, and acute physiology and chronic health evaluation II (APACHE II). Results: Out of 291 eligible patients, 250 were enrolled with a mean age (± standard deviation, SD) of 60.23 years (± 14.49), and gender ratio of 1.72:1, male: female. On admission to the ICU, 72.8% had been transferred from the emergency room (ER), 20.8% from other hospital departments and 6.4% from other hospitals. The percentage of patients taking antibiotics prior to ICU admission was 64.8% (n = 162). Antibiotic therapies were modified for 35.6% (n = 89) of patients upon ICU admission and for 49.2% (n = 123) of patients during their ICU stay. The sequential organ failure assessment (SOFA) score on the day of antibiotic change was higher than the SOFA score on ICU admission (P = 0.002). Conclusions: Antibiotic therapies are very common in the ICU pre-admission period, as are antibiotic regimen changes post-admission. The SOFA score is a potentially reliable tool for assessing the need to alter antibiotic therapy.