Does Perioperative Blood Transfusion In¯uence Long-Term Prognosis of Gastric Cancer? F. S Â ANCHEZ-BUENO, MD, PhD, J.A. GARCIA-MARCILLA, MD, PhD, J.M. P Â EREZ-ABAD, MD, R. VICENTE, MD, F. ARANDA, MD, J.A. LUJAN, MD, PhD, and P. PARRILLA, MD, PhD We analyze d the in¯uence of packed red blood cell (PRBC) transfusions on the prognosis of 163 patients with gastric adenocarcinoma undergoing subtotal gastrectomy with a curative intention. Over a period of 15 years, our department admitted 505 patients with gastric adenocarcinoma, with curative subtotal gastrectomy being performed in 167 cases. Mean age was 62.2 years (range: 30 ±87); there was a predominance of males (104 cases; 63.8% ). Excluding the four patients who died in the immediate postope rative period (®rst 30 days), the remaining 163 were reviewed twice yearly in our department until either they died or the study ended. Follow-up averaged 49.5 months, with a median of 36 months. Sixty-nine (42.3% ) of the 163 patients received transfusions of PRBC. On correlating the variable s with the transfusion, we found a statistical signi®cance only between the rate of transfusion and patient age over 63 years ( P , 0.01), with an evolution time of less than three months ( P , 0.05) and in tumors of .4 cm ( P , 0.05). The ®ve-year survival rate of the nontransfusion patients was 56.9% and of the transfusion patients 40%, with statistically signi®cant differ- ences ( P 5 0.0132). On studying patients according to tumor stage, we found that blood transfusion had a statistically signi®cant in¯uence on prognosis only in patients with tumor stage III ( P 5 0.0051). In the univariate analysis of the remaining variables collected, the existence of abdominal tumor ( P 5 0.0307), tumor size ( P 5 0.00001) , degree of involve ment of the gastric wall ( P 5 0.00001), lymph node involvement ( P 5 0.00001) and tumor stage ( P 5 0.00001) revealed a statistically signi®cant in¯uence on prognosis. If we apply Cox’s regression model to the variables that in the univariate analysis had a statistically signi®cant in¯uence on prognosis, we found that only tumor size and stage were independent predictors of survival. In our experience, PRBC transfusion does not in¯uence the long-term survival of patients with resected gastric adenocarcinoma. KEY WORDS: stomach; gastric adenocarcinoma; blood transfusion; subtotal gastrectomy; prognosis. In 1982 Burrows and Tartter (1) were the ®rst to report that blood transfusion (BT) favored the recur- rence of resected colorectal cancer. Since then other authors (2±5) have shown that BT has a deleterious effect on various immune mechanisms of the host, facilitating the spread of neoplastic cells. However, there are few publications related to gastric cancer, and those that exist are contradictory, because some authors (6, 7) ®nd no correlation between BT and survival in resected gastric cancer, whereas others (8, 9) consider BT to be a predictive factor of a poor prognosis. Manuscript received January 8, 1997; revised manuscript re- ceived June 26, 1997; accepted June 27, 1997. From the Department of General Surgery, ªVirgen de la Arrixacaº University Hospital, Murcia, Spain. Address for reprint requests: Prof. Dr. Francisco Sa Ânchez Bueno, Servicio de CirugõÂa, 3 a planta, Hospital Universitario ªVir- gen de la Arrixaca,º 30120, Murcia, Spain. Digestive Diseases and Sciences, Vol. 42, No. 10 (October 1997), pp. 2072±2076 2072 Digestive Diseases and Sciences, Vol. 42, No. 10 (October 1997) 0163-2116/97/1000-2072$12.50/0 Ñ 1997 Plenum Publishing Corporation