A retrospective analysis of the records of 60 patients who underwent pericardiectomy over a 10-year period (1980 to 1990) at The Johns Hopkins Hospital was performed. Indications for operation were effusive disease in 24 patients and constriction in 36 patients. Six patients (10%) with pericardial effusion had pain as the primary symptom necessitating intervention. The operative approach for pericardiectomy was median sternotomy in 52 patients (4 patients required cardiopulmonary bypass) and left anterior thoracotomy in 8 patients. Nine patients (5 with constriction and 4 with effusion) with a prior limited pericardial procedure required formal pericardiectomy. The operative mortality rate for pericardial effusion and constriction was 4.2% and 5.6%, respectively. Follow-up (median follow-up, 56.9 +/- 38.2 months) was obtained on 56 patients (93.3%). Actuarial survival at 1 year, 5 years, and 10 years for all patients was 82.1% +/- 5.1%, 71.7% +/- 6.7%, and 59.8% +/- 12.2%, respectively. A Cox proportional hazards regression analysis was performed using 20 clinical variables. A history of malignancy, previous pericardial procedure, and preoperative New York Heart Association class IV were found to be predictors of poor survival. All patients who underwent operation primarily for effusion with associated pain are alive and have improved functional capacity without steroid use. We conclude that pericardiectomy can be performed with low mortality and can result in good long-term survival and improved functional capacity. Patients who are seen primarily with pain refractory to steroid therapy can be relieved of symptoms with operation.