Recent studies suggest that pretransplant blood transfusions prolong kidney graft survival by nonspecific immune suppression. Since immune suppression in patients with cancer is associated with early recurrence and poor prognosis, we studied the relationship of perioperative blood transfusion to recurrence of Stage I (subset N0) lung cancer. Technique-related variables were minimized by limiting the study to one surgeon (P. A. K.). Life-table and Cox proportional hazards analysis of age, sex, tumor size, histopathology, admission and discharge hematocrit values, estimated operative blood loss, duration of operation, extent of resection, anesthetic agents, and blood transfusion revealed two statistically significant prognostic factors: extent of resection (p = 0.0056) and use or non-use of transfusions (p = 0.0283). The cumulative 5 year disease-free survival rate of the 15 patients undergoing pneumonectomy was 42% compared to 75% for 150 patients undergoing lobectomy and lesser procedures. Transfused patients had lower disease-free rates within 5 years than nontransfused patients (62% versus 76%, p = 0.0132). Survival advantage was also noted in nontransfused patients after deleting patients subjected to pneumonectomy. These results indicate that perioperative transfusion in patients with lung cancer undergoing resection accelerates the appearance of recurrent or metastatic cancer. This supports the findings of previous studies of breast and colon cancer that blood transfusion is detrimental to the patient with cancer.