Inflammatory responses play major roles in the development of acute lung injury following lung cancer surgery. The authors tested the hypothesis that thoracic epidural anesthesia (TEA) during surgery could attenuate both systemic and local inflammatory cytokine productions in patients undergoing lung cancer surgery. A prospective randomized controlled trial. At Keio University Hospital, Tokyo, Japan. Patients scheduled for lung cancer surgery. Sixty patients were randomly allocated into two groups (n = 30 each group): the epidural group (group E), in which anesthesia was maintained with propofol, fentanyl, rocuronium, and epidural anesthesia with 0.25% levobupivacaine; or the remifentanil group (group R), in which a remifentanil infusion was used as a potent analgesia instead of epidural anesthesia. The lung epithelial lining fluid (ELF) and blood sampling were collected prior to one-lung ventilation (OLV) initiation (T1) and at 30 minutes after the end of OLV (T2). The concentrations of tumor necrosis factor (TNF)-α, interleukin (IL)-6, and IL-10 in the ELF at T2 were increased significantly compared with those at T1 in both groups. The ELF concentration of IL-6 in group E was significantly lower than that in group R at T2 (median [interquartile range]: 39.7 [13.8-80.2] versus 76.1 [44.9-138.2], p = 0.008). Plasma IL-6 concentrations at T2, which increased in comparison to that at T1, were not significantly different between the two groups. The plasma concentrations of TNF-α did not change in both groups. This randomized clinical trial suggested that TEA could attenuate local inflammatory responses in the lungs during lung cancer surgery.