To assess the incidence of a specific etiology and the role of methods for specific etiologic diagnosis in patients with primary acute pericarditis, we studied 100 patients with primary acute pericarditis consecutively admitted to our hospital between 1991 and 1993. A general diagnostic protocol was performed in all patients, whereas only pericardiocentesis was performed in patients with clinical cardiac tamponade or an unfavorable course with anti-inflammatory drugs. Surgical drainage and pericardial biopsy was performed in patients with tamponade relapse. A specific etiology was discovered in 22 patients (22%) (neoplasms in 7, tuberculosis in 4, other infections in 3, collagen diseases in 3, thyroid disorders in 4, and dissecting aortic aneurysm in 1). The general diagnostic protocol led to a specific diagnosis in 15 patients (68% of all patients with specific acute pericarditis) and pericardiocentesis in the other 7 patients (32%). The role of a diagnostic protocol, therapeutic pericardiocentesis, and diagnostic pericardiocentesis was similar and complementary. Pericardial biopsy results were negative in the 5 patients in whom it was performed. Cardiac tamponade and an unfavorable clinical outcome were significantly (p < 0.001) associated with the finding of a specific etiology; when both features were combined, sensitivity was 86% and specificity 85%, positive predictive value 63% and negative predictive value 96%. We conclude that the specific etiology in patients with primary acute pericarditis is about 20% to 25%, and that about 90% of these specific cases can be discovered by using the described systematic diagnostic protocol only in patients with an unfavorable outcome (cardiac tamponade or poor clinical course).