Prognostic significance of perioperative blood transfusions in resectable thoracic esophageal cancer

OBJECTIVE: The perioperative blood transfusions have been associated with t umor recurrence and decreased survival in various types of alimentary tract cancer. There exist, however, contradictory studies showing no relationshi p between blood transfusions and survival. For patients with esophageal can cer, only one report suggested that blood transfusions did not by itself de crease the chance of cure after esophagectomy. METHODS: Among 235 patients with primary squamous cell carcinoma of the tho racic esophagus between December 1979 and March 1998, 143 patients (60.9%) underwent esophagectomy with curative intent (RO). To exclude the effects o f surgery-related postoperative complications, 14 patients who died within 90 days during the hospital stay were excluded. Thus, clinicopathological c haracteristics and prognostic factors were retrospectively investigated bet ween patients with no or few transfusions (less than or equal to 2 units) ( n = 58), and much transfused patients (greater than or equal to 3 units) (n = 71). RESULTS: Sixty-three patients are alive and free of cancer, and 66 patients are dead. A total of 98 patients (76%) received blood transfusions, wherea s 31 patients (24%) had no transfusion. The amount of blood transfused was 1 or 2 units in 27 patients (27.6%), 3 or 4 units in 33 (33.7%), 5 or 6 uni ts in 20 (20.4%), and greater than or equal to 7 units in 18 (18.4%). The 5 -yr survival rate for patients with no or few transfusions was 69%, whereas that for much transfused patients was 31.7% (p < 0.0001). The much transfu sed patients had more prominent ulcerative tumor, longer time of operation, more estimated blood loss, and more marked blood vessel invasion than the group with no or few transfusions. The factors influencing survival rate we re tumor location, Borrmann classification, size of tumor, depth of invasio n, number of lymph node metastases, time of operation, amount of blood tran sfusions, lymph vessel invasion, and blood vessel invasion. Among those nin e significant variables verified by univariate analysis, independent progno stic factors for survival determined by multivariate analysis were number o f lymph node metastasis (0 or 1 vs greater than or equal to 2, p < 0.0001), amount of blood transfusions (less than or equal to 2 units vs greater tha n or equal to 3 units, p < 0.0001), and blood vessel invasion (marked vs no n-marked, p 0.0207). CONCLUSIONS: There is an association between high amount of blood transfusi ons and decreased survival for patients with resectable esophageal cancer. To improve the prognosis, surgeons must be careful to reduce blood loss dur ing esophagectomy with extensive lymph node dissection and subsequently mus t minimize blood transfusions. (Am J Gastroenterol 1999;94:757-765. (C) 199 9 by Am. Coll. of Gastroenterology).