BACKGROUND AND OBJECTIVES: The identification of a specific etiology of effusive pericardial disease is difficult because of the limited yield of cytologic and microbiologic pericardial fluid analysis. We performed retrospective study to find out whether pericardial biopsy was superior to pericardial fluid analysis in search of the etiology of pericardial effusion. MATERIALS AND METHOD: We reviewed 76 cases of moderate to severe pericardial effusion on which we performed surgical pericardial biopsy from Sep. 1986 to Sep. 1996. The results of pericardial fluid analysis, clinical manifestation, pericardial biopsy were compared retrospectively. RESULTS: 1)Clinical diagnosis of pericardial effusion were as follow:neoplastic disease (7.9%), tuberculosis (72.4%), constrictive pericarditis (17.1%), and others (2.6%). 2)By the percutaneous pericardial biopsy, we confirmed 19 cases (25%). Etiology of 4 cases (5.3%) were malignancy and 15 cases (19.7%) tuberculosis. Fifteen out of 76 patients who were diagnosed by biopsy as tuberculous pericarditis and 28 patients who were suspected as tuberculous pericarditis clinically were treated with antituberculous medications. Ten patients (66.7%) of pathologically diagnosed patients and 18 patients (69.2%) of clinically diagnosed patients showed complete resolution of pericarditis. CONCLUSION: By pericardial biopsy, we only confirmed 19 cases (25.0%). It means that pericardial biopsy is not superior to pericardial fluid analysis in searching of etiology of pericardial effusion. Moreover, it is not sufficient for final diagnosis of pericardial effusion.