Cardiac surgery after mediastinal radiation: Extent of exposure influences outcome

Objectives: Mediastinal radiation for thoracic malignancies uses multiple treatment fields and doses. We investigated whether more extensive radiation exposure is associated with more hospital complications and worse survival after cardiac surgery. Methods: From January 2000 to January 2005, 230 patients underwent cardiac surgery after 3 levels of mediastinal radiation: extensive (Hodgkin disease, thymoma, and testicular cancer; n = 70), variable (eg, non-Hodgkin lymphoma and lung cancer; n = 35); and tangential (breast cancer; n = 125). Hospital complications were recorded prospectively, and time-related survival was assessed by patient follow-up (mean follow-up, 2.2 +/- 1.4 years). Results: Patients receiving extensive exposure were youngest (51 vs 64 vs 72 years), with the longest radiation-to-operation interval (25 vs 13 vs 14 years), and had the most diastolic dysfunction, left main stenosis of greater than 70% (21% vs 9% vs 8%), and aortic regurgitation (79% vs 54% vs 50%). Patients receiving extensive and variable exposure had the poorest pulmonary function (percent predicted forced expiratory volume in 1 second, 57% vs 54% vs 67%; percent predicted forced vital capacity, 56% vs 63% vs 66%). All groups received a similar mix of cardiac procedures. Hospital deaths (13% vs 8.6% vs 2.4%) and respiratory complications (24% vs 20% vs 9.6%) were higher after more extensive radiation, and survival was poorer (4-year survival, 64% vs 57% vs 80%) than for patients receiving tangential radiation exposure, and it deviated more from expected matched-population life tables. Conclusions: Among patients undergoing cardiac surgery after thoracic radiation, radiation exposure is heterogeneous, and therefore these patients cannot be managed and assessed as a single uniform cohort. Extensively irradiated patients are more likely to develop radiation heart disease, which increases perioperative morbidity and decreases short- and long-term survival.