Objective: To determine the adequate frequency of empirical antibiotic treatment in an intensive care unit in Colombia. Methodology: Observational, analytic study. Data was collected from the clinical and microbiological history of the intensive care unit (ICU) at Clinica de Tunja. Data was collected from the empirical antibiotic regimen, antibiograms, and percentage of adjustment according to culture results. The percentage of adequate empirical therapy was determined based on the cultures and antibiograms. Results: We studied 198 cases of infection in ICU between July 2008 and September 2011. Seventy-five percent of infections was documented by microbiological media, which corresponds to 148 patients with 165 microorganisms. In 94 of 148 patients with infection documented by cultures, empirical treatment was adequate (64%). In cases of infection not documented by cultures, empirical therapy was adequate in 36 of 50 patients (72%). In 128 (65%) of cases, empirical treatment was changed. Empirical treatment was adequate in 100% of cases that received imipenem, piperacillin/tazobactam and amikacin, compared to 68% when imipenem/amikacin was used (p <0.001) and 29% when only imipenem was used (p <0.001). Conclusions: The percentage of adequate empirical therapy was low (65.6%) and increased with the combination of three antimicrobials in populations with high probability of infection by multi-resistant germs.