OBJECTIVE--To identify features associated with success or failure of aspiration of pericardial effusion. METHOD--A retrospective analysis of 36 drainage procedures in 30 patients with pericardial effusion was performed using patient records and echocardiograms. RESULTS--Unsuccessful aspiration was associated with pericardial loculation but not with the seniority of the operator or the size and position of the effusion. Pericardiocentesis relieved symptoms of breathlessness in 21 of 26 patients who had a pericardial effusion suspected of causing dyspnoea. These 21 patients had few clinical or echocardiographic signs of classic tamponade. CONCLUSION--The paucity of abnormal physical or echocardiographic signs of tamponade in breathless patients with pericardial effusion does not exclude symptomatic benefit being derived from pericardiocentesis. Pericardial aspiration is safe in appropriate hands, although aspiration of loculated effusions may not be as successful as aspiration of non-loculated effusions.