We investigated the relationship between surgical stress and tumor metastasis. The excessive surgical stress of a thoracolaparotomy enhanced tumor metastasis remarkably in an experimental model. We would like to propose that this phenomenon be termed "surgical oncotaxis". This effect has previously been attributed to some mechanisms of immunosuppression, excessive secretion of corticoids, and active oxygen production of granulocytes. An increase in lipid peroxide (LPO) in the liver was observed after a thoracolaparotomy, but a strong radical scavenger of a DL-alpha-tocopherol-L-ascorbic acid 2-0-phosphate diester (EPC-K1) restrained LPO levels in the liver and the effect of tumor metastasis in parallel. As clinical strategies for restraining the surgical oncotaxis, the control of any cytokine storm after surgery and/or the scavenging of active oxygen appears to be possible and hopeful, since it might be intermediated by cytokine. When pre-administration findings for EPC-K1 and methylpredonisolone were compared, EPC-K1 was found to be more suitable for restraining surgical oncotaxis, because serum LPO was only controlled with EPC-K1. The cytokine storm which occurs after surgery is augmented by a second stimulation, such as the administration of lipopolysaccharide, and no drug could control this well experimentally. Postoperative complications are a clinical model of a second stimulation (a so-called second attack). Our data showed the prognosis of a group with complications to be worse than that of a group without them even though no difference existed in the background of the esophageal cancer patients studied. Based on these results, safe surgery and the choice of minimally invasive surgery are the best ways to control surgical oncotaxis. Following a major surgical procedure, such as a thoracolaparotomy, the use of corticoids and/or radical scavengers can contribute to restraining surgical oncotaxis.