In cases of malignant pericardial effusion, surgical subxiphoid biopsy sometimes fails to prove malignancy. To assess the usefulness of pericardioscopy, which allows an endoscopic investigation of the pericardial cavity, this technique was systematically performed during surgical drainage procedures that were performed on 40 patients who had pericardial effusions of suspected malignant origin. Twenty-six patients had a history of neoplasm, 10 had a history of hematologic malignancy, and four had recent tumors or lymphadenopathies that were suspected to be of malignant origin. Classical tests that are usually performed during a conventional surgical drainage procedure (fluid studies and subxiphoid biopsy) were combined with direct visualization of the pericardial surfaces and guided biopsies of suspicious areas. The follow-up period after pericardioscopy was at least 12 months. Two early deaths occurred after pericardioscopy, but no death was directly related to the endoscopy. According to all of the tests that were performed, diagnoses were malignant pericardial effusion in 15 of 40 patients (group I, 37#) and nonmalignant pericardial effusion in 25 of 40 patients (group II, 73%). In 3 of 13 patients (23%) in group I, the diagnosis was obtained only by pericardioscopy (results of cytologic studies and subxiphoid biopsy were negative). In two patients in group I, pericardioscopy could not be completed, but the diagnosis of malignant pericardial effusion was obtained by pericardiocentesis. In group II, effusion was considered to be postradiation pericarditis in five cases, infectious pericarditis in three cases (bacterial in one and tuberculous in two), hemopericardium induced by coagulation disturbances in three cases, and idiopathic pericarditis in 14 cases. Mid- and long-term follow-up (mean duration, 35 months; range, 12 to 72 months) showed that in group I the median survival time was 42 days, whereas in group II it was 1 year. The difference between the two life curves was significant ( p < 0.05). Thus pericardioscopy appears to be a technique that does not increase the risk of the surgical drainage procedure. Its main benefit is a greater diagnostic sensitivity as a result of direct visualization of the pericardial surfaces and guided biopsies. It should be available at the time that the surgical drainage procedure is performed, since the diagnosis of malignant pericardial effusion has significant prognostic consequences.