INTRODUCTION: Blood stream infections (BSIs) are responsible for significant morbidity and mortality. Numerous studies have pointed to the importance of early and prompt institution of empirical antimicrobial therapy in reducing morbidity and mortality in BSI. The antibiotic spectrum must be narrowed as soon as possible, considering the clinical condition of the patient, the pathogens identified in cultures and the sensitivity profile obtained from the antibiogram. When no evidence of bacterial infection is present, antibiotic therapy must be suspended. However, it is often observed that the clinician does not adhere to the guideline and continues with the same empirical treatment. MATERIALS AND METHODS: The study was conducted at Department of Microbiology, in a tertiary care hospital from April 2017 to September 2017. Two hundred patients with microbiologically documented BSIs were included in the study. They were followed up to find the appropriateness of change in empirical treatment carried out according to culture sensitivity report, and antimicrobial consumption was also calculated. RESULTS: We observed that there was an increased use of antipseudomonal penicillins plus beta-lactamase inhibitors, amikacin and carbapenems which was due to higher prevalence of multidrug-resistant Gram-negative bacilli among blood culture isolated. We also found that in most of the patients, the empirical treatment was inappropriately modified at day four after availability of culture reports, i.e., inappropriately escalated or de-escalated or continued while there was no indication to do so. CONCLUSION: Treatment inappropriate group was associated with higher treatment failure as compared to treatment appropriate group. Studies of the other factors associated with inappropriate treatment such as changes in resistance patterns, antimicrobial-related adverse effect and of the long-term clinical outcomes are warranted.