Effects of albumin and crystalloid priming strategies on red blood cell transfusions in on-pump cardiac surgery: a network meta-analysis.

Background: In on-pump cardiac surgery, the albumin priming strategy could maintain colloid osmotic pressure better than crystalloid solutions and reduce excessive perioperative fluid balance. However, a high-quality meta-analysis is required to compare the safety of these approaches in perioperative red blood cell (RBC) transfusions. Owing to limited direct evidence, we conducted a network meta-analysis (NMA) to increase the pool of studies and provide indirect evidence.

Methods: The pre-defined primary outcomes were intraoperative and the first 24 h postoperative RBC transfusion volume in units. The pre-defined secondary outcome was postoperative blood loss (the first 24 h). We reviewed all randomized controlled trials comparing albumin, crystalloid, and artificial colloid priming strategies. Studies that only displayed pre-defined outcomes could be included. A pairwise meta-analysis was performed on studies that directly compared the pre-defined outcomes between albumin and crystalloids. Additionally, a random-effects network meta-analysis (NMA) model was employed to generate indirect evidence for the pre-defined outcomes between albumin and crystalloids.

Results: The literature search identified 830 studies,10 of which were included in the final analysis. Direct meta-analysis indicated that crystalloid priming significantly decreased total perioperative RBC transfusions (MD: -0.68U; 95%CI: -1.26, -0.09U; P = 0.02) and intraoperative RBC transfusions (MD: -0.20U; 95%CI: -0.39, -0.01U; P = 0.03) compared to albumin. Postoperative RBC transfusions showed a decreasing trend in the crystalloid group; however, the difference was not statistically significant. (MD: -0.16U; 95%CI: -0.45, 0.14U; P = 0.30). After including indirect evidence, the NMA results continued to demonstrate a higher RBC receiving with the albumin priming strategy compared to crystalloids, although the differences did not reach statistical significance. For postoperative blood loss, direct evidence showed no significant differences between albumin and crystalloid priming strategies. However, NMA evidence displayed that albumin exist higher probability of reducing postoperative blood loss than crystalloid.

Conclusion: Both direct and NMA evidence indicated that the albumin priming strategy resulted in more perioperative RBC transfusions than crystalloids. Considering the additional blood management burden, the application of an albumin-priming strategy in on-pump cardiac surgery still needs more consideration.