Objective To assess the effect on survival of macrolides or fluoroquinolones in intubated patients admitted to the intensive care unit (ICU) with severe community-acquired pneumonia (severe CAP). Methods Prospective, observational cohort, multicenter study conducted in 27 ICUs of 9 European countries. Two hundred eighteen consecutive patients requiring invasive mechanical ventilation for an admission diagnosis of CAP were recruited. Results Severe sepsis and septic shock were present in 165 (75.7%) patients. Microbiological documentation was obtained in 102 (46.8%) patients. ICU mortality was 37.6% ( n  = 82). Non-survivors were older (58.6 ± 16.1 vs. 63.4 ± 16.7 years, P  < 0.05) and presented a higher score on the simplified Acute Physiology Score II at admission (45.6 ± 15.4 vs. 50.8 ± 17.5, P  < 0.05). Monotherapy was given in 43 (19.7%) and combination therapy in 175 (80.3%) patients. Empirical antibiotic therapy was in accordance with the 2007 Infectious Diseases Society of America (IDSA)/American Thoracic Society (ATS) guidelines in 100 (45.9%) patients (macrolides in 46 patients and fluoroquinolones in 54). In this cohort, a Cox regression analysis adjusted by severity identified that macrolide use was associated with lower ICU mortality (hazard ratio, HR 0.48, confidence intervals, 95% CI 0.23–0.97, P  = 0.04) when compared to the use of fluoroquinolones. When more severe patients presenting severe sepsis and septic shock were analyzed ( n  = 92), similar results were obtained (HR 0.44, 95% CI 0.20–0.95, P  = 0.03). Conclusions Patients with severe community-acquired pneumonia had a low adherence with the 2007 IDSA/ATS guidelines. Combination therapy with macrolides should be preferred in intubated patients with severe CAP.