Background: Perioperative transfusion (POT) is a negative prognostic factor for outcome in cancer patients undergoing surgery, and splenectomy (SPL) has been suggested to help obviate this negative effect. However, the prognostic independence among related factors and clinical significance of the interaction remains elusive.
Patients and methods: Between January 1, 1990 and January 31, 2005, a total of 200 patients with proximal gastric cancer classified as having stage IB to III disease after curative resection were enrolled at Kitasato University Hospital in a retrospective manner, according to a previous report by Weitz et al.
Findings: POT was reconfirmed to be a potent univariate prognostic predictor in this model (p < 0.0001). The tendency was further augmented in patients without SPL, suggesting that SPL can counteract the negative aspects of POT. A multivariate analysis including related factors revealed that POT was one of the most potent independent prognostic factors, as were stage and age. Stratification analysis following the JCGC staging system revealed that the outstanding prognostic difference according to POT was recognized in patients anticipated to be cured, such as those with stage IB (p = 0.012) or II (p = 0.003) but not stage III. SPL could dramatically improve the prognosis with POT in stage IB patients but modestly if at all in those with stage II or III.
Interpretation: Our findings support the need for a clinical alert against POT throughout the course of cancer treatment, especially for curable disease; thus, POT should be carefully considered from an oncologic point of view. Appropriate strategies to minimize POT should be urgently developed.