Obtaining diagnostic microbiological cultures before initiating empirical antimicrobial therapy is part of the diagnostic work-up of intensive care patients with a clinical suspicion of infection. However, it is unknown to what extent these cultures provide a microbiological cause of infection and to what extent antimicrobial therapy is influenced. During a 6-month period, all episodes of suspected clinical infection were analysed and categorised as non-microbiologically proven infection (non-MPI) or MPI. Effects of culture results on antibiotic therapy were analysed for episodes of respiratory tract infection. Invasive diagnostic techniques were not routinely used for diagnosis of respiratory tract infections. Among 212 patients admitted, 147 episodes of clinical suspicion of infection were recorded (104 for respiratory tract infection) and 1147 microbiological cultures were obtained (0.64 culture per patient day). Antibiotics were administered on 1111 (62%) of 1803 patients days. Of the respiratory tract infections, 571 cultures resulted in 49 (47%) MPI. Cover with empirical antibiotics was inappropriate in 7 of 104 cases (8%) of respiratory infections. In 12 cases (11.5%) empirical therapy could have been changed based on culture results. Negative cultures were never followed by cessation of therapy, but the duration of treatment was significantly shorter for non-MPI. Forty-seven percent of respiratory tract infections were microbiologically confirmed and, based on culture results, empirical antimicrobial therapy could have been influenced in 11.5% of cases of respiratory tract infections. These findings provide aspects to evaluate and improve the diagnostic work-up of infections in the ICU.