In pleural infection, it has been recommended that Gram stain and cultures should be obtained on a routine basis. However, this recommendation has not been tested prospectively. We evaluated the yield of microbiological studies in 259 patients with parapneumonic pleural effusion. Microbiological studies were positive on the pleural fluid of 50 patients (19.3%). In 48 of the 50 patients with positive microbiological results (96%), the need for pleural drainage was correctly predicted by pleural fluid parameters. There were no differences in hospital stay (9.5+/-2.5 days versus 9.9+/-3.2 days, P=0.68) or in mortality (2 deaths in each group, P=0.58) between the group of patients in which antibiotic treatment was changed according to microbiological results and the group of patients in which it is not. In conclusion, this study demonstrates that, at least in our institution, routine microbial investigation of pleural fluid adds very little to the standard management of parapneumonic effusions.