RESUMO
OBJECTIVE: To examine the influence of perioperative blood transfusions on perioperative outcomes and late survival. BACKGROUND: Perioperative blood transfusion has been reported to have a negative impact on perioperative morbidity but its long-term effect on survival is unknown. The purpose of this study was to evaluate the effects of perioperative transfusion on perioperative outcomes and survival. METHODS: We studied 12,345 surgical procedures from Veteran Administration Surgical Quality Improvement Program database from July 1998 through 2010. Patients with transfusion were compared with a severity-matched control group. We performed the Fisher exact test for comparison of categorical values and Wilcoxon rank sum test for continuous values. Multivariate regression was used to eliminate other confounding factors. The predictive value of multivariate risk model was tested with receiver-operator curves. Patients were matched using an optimal 1:1 digit-matching algorithm. All analyses were performed with NCSS-2007 version 1-12. P < 0.05 was considered statistically significant. RESULTS: The 848 patients who received perioperative transfusions had higher unadjusted rates of mortality and decreased long-term survival. The odds ratio (OR) for 10 years mortality in transfused group was 2.92 and after adjusting for preoperative risk factors decreased to 1.40 (P < 0.01). However, after data were filtered for any perioperative complications, such an association was seen before, OR = 2.05 (P = 0.006), and was lost after propensity matching, OR = 1.19 (P = 0.35). CONCLUSIONS: After filtering out perioperative complications and adjusting for preoperative morbidity, our final analysis did not reveal an increased long-term mortality. We conclude that transfusion may reduce long-term survival through its effects on perioperative complications.