BACKGROUND: Symptomatic pericardial effusion in patients with cancer may lead to a life-threatening event that requires diligent treatment, but the best surgical treatment is still controversial. The purpose of this study was to identify predictors of survival for patients with solid malignancies and symptomatic pericardial effusion, which might help to select the best surgical treatment for each patient. METHODS: We retrospectively analyzed 47 patients with solid malignancies concomitant with symptomatic pericardial effusion who underwent surgery between 1994 and 2004. Overall survival was calculated from date of surgery, and prognostic importance of clinical and pathological variables was assessed. RESULTS: The most common primary sites of disease were breast (46.8%) and lung (25.6%). Initial pericardiocentesis were performed in 29 patients; median volume of fluid drained was 480 mL. Median interval from the diagnosis of primary cancer to the development of pericardial effusion (pericardial effusion-free interval) was 34.8 months. Definitive surgical treatment was performed in 43 patients, as follows: subxiphoid pericardial window (n = 21); thoracotomy and pleuropericardial window (n = 10); pericardiodesis (n = 8); and videothoracoscopic pleuropericardial window (n = 4). Pericardiocentesis was the only procedure in four patients. Median follow-up was 2.9 months. Median overall survival was 3.7 months. Pericardial effusion-free interval longer than 35 months and more than 480 mL of fluid drained at initial pericardiocentesis were determinants of better survival. CONCLUSIONS: Pericardial window and pericardiodesis seem to be safe and efficacious in treating effusion of the pericardium. Pericardial effusion-free interval and volume drained at initial pericardiocentesis are determinants of outcome.