Pneumonia is an important cause of morbidity and mortality among children throughout the world. Vaccines are available for some organisms, but they are underutilized and/or still in development. To evaluate the potential impact of vaccines, we review studies in which the etiology of childhood community-acquired pneumonia was recorded. In North America and Europe (9 studies), the etiology of pneumonia was established in 62% of studied children (range 43%-88%) by use of noninvasive specific methods for microbiologic diagnosis. The most often identified agents were S. pneumoniae (22%), respiratory syncytial virus (RSV) (20%), Haemophilus influenzae (7%), and Mycoplasma pneumoniae (15%). In Africa and South America (8 studies), bacteria were recovered from 56% (range 32%-68%) of severely ill children studied by lung aspirate. The most often isolated bacteria were Streptococcus pneumoniae (33%) and Haemophilus influenzae (21%). A high percentage of H. influenzae strains were not serotype b. Throughout the world, children requiring hospitalization were most likely to have infection caused by pneumococcus, H. influenzae or RSV. Out patients also had Mycoplasma pneumoniae. Countries in Africa and Asia recorded 2 to 10 times more children with pneumonia (7 to 40/100 annually) than in the USA. Widespread use of pneumococcal and H. influenzae type b conjugate vaccines could reduce the frequency of childhood pneumonia by one-third. Further reduction will require development of non-type b H. influenzae, RSV and M. pneumoniae vaccines. This could result in a > 50% reduction of pneumonia in children. This goal should be sought and achieved as soon as possible.