Community-acquired pneumonia (CAP) is common, costly, and clinically serious. Several national and international practice guidelines have been developed to promote more appropriate, cost-effective care for patients with CAP. This article compares and contrasts eight international practice guidelines for the management of CAP, describes the extent to which recommendations are reflected in practice, and proposes explanations for non-adherence to guidelines. We found consistency in recommendations across all the guidelines for the management of patients with CAP requiring intensive care. In this setting, all guidelines recommend chest radiography, sputum Gram stain and culture, blood cultures, testing for Legionella pneumophila, and timely administration of antibiotics active against both typical (i.e. Streptococcus pneumoniae, Hemophilus influenzae) and atypical organisms (i.e. Legionella spp., Mycoplasma pneumoniae, and Chlamydia pneumoniae). Recommendations for the management of the average inpatient with pneumonia were more variable, with the greatest differences between the North American and European guidelines. The North American guidelines (in contrast to European ones), recommended empiric treatment of typical and atypical organisms in all inpatients. There were also differences in policies regarding the necessity of chest radiography, sputum studies, and serologic testing. Some guidelines explicitly embrace the use of prediction rules to inform the decision to hospitalize, while others do not. Some of these admission decision algorithms focus on identifying low risk patients, while others are most concerned with high risk patients. There was also considerable variation in the specificity and operationalization of clinical criteria for switching from parenteral to oral antibiotics or judging appropriateness for discharge. Many recommendations for key management decisions tended to lack explicit, objective, and actionable criteria that could be easily implemented in real world practice. Review of the pneumonia literature revealed that physician performance of guideline-recommended best practices is often suboptimal. Administration of timely antibiotics (< or =8 hours of presentation) and use of first-line antibiotics occurred in 75-85% and 18-79% of cases, respectively. Collection of blood cultures within 24 hours of presentation and prior to administration of antibiotics was achieved in 69-83% and 63-82% of cases, respectively. Screening the eligibility of CAP patients for hospital-based pneumococcal and influenza vaccination occurred on average in 11 and 14% of hospitalizations, respectively, in the US. Lack of awareness of guidelines, conflicting advice among them, and lack of specific, objective, actionable recommendations most likely contribute to nonadherence to CAP guidelines. Increased attention to these factors will be needed if professional society practice guidelines are to fulfill their promise as tools for improving the quality and outcomes of care for patients with pneumonia.