Pericardial effusion requires attention because of its underlying pathology and its potential for producing hemodynamic compromise. It was our hypothesis that the etiology of and danger from large pericardial effusions are different at different ages. We reviewed the clinical data on 44 consecutive patients seen between 1986 and 1993. All (age range 0-19 years) had a large pericardial effusion identified by echocardiography, and 38 had fluid available for analysis, pericardiocentesis in these cases being undertaken on the basis of the clinical decision of the attending physician. History, physical examination, chest x-ray, electrocardiogram, and echocardiographic Doppler findings were evaluated retrospectively. Of the patients, 17 presented with tamponade. Of the 14 patients aged less than one year, 12 had tamponade (85%), whereas in the 30 patients older than one year, only five (16%) had tamponade. Surgical drainage was needed in seven patients. Typical physical and electrocardiographic findings (distant heart sounds, pericardial friction rub, decreased QRS voltage, and ST-T wave abnormalities) were present in less than half the patients. Etiologies of the effusion included infection in 10 (four less than one year old, and six greater than one year of age). Among these, two unusual pathogens were detected: Mycobacterium avium-intracellulaire and Candida tropicalis. In eight patients, the effusion was due to trauma (six less than one year old, and two older than one year), and in 11 it was due to post-pericardiotomy syndrome (two less than one year of age and nine older than one year). Malignancy was the cause in six (all older than one year of age), radiation in two, and collagen vascular disease, drug induced, and uremia in one each, all older than one year. Only three patients had idiopathic effusions (6%). These data suggest that, unlike adults, an etiologycan be found in most children with large pericardial effusions. Those less than one year of age are much more likely to present in cardiac tamponade.