The Predictive Role of Intraoperative Blood Transfusion Components in the Prognosis of Heart Transplantation

Abstract

Purpose: To evaluate the influence of transfusion amount of blood components on the prognosis of patients after heart transplantation (HTx).

Methods: From 1 January 2015 to 31 December 2020, 568 patients underwent HTx in our institute. A total of 416 recipients with complete datasets were enrolled in the study for final statistical analysis according to the inclusion criteria. The optimal cut-off values for intraoperative transfusion of red blood cell (RBC), platelet, and plasma were determined with receiver operating curve analysis. Univariate and multivariate Cox regression analyses were applied to compare baseline data of patients divided by the transfusion amounts of RBC, platelet, and plasma. Propensity score matching was used to enable the direct comparison of outcomes.

Results: The Kaplan-Meier analysis revealed that transfusion amounts of RBC and plasma were independently associated with overall mortality, increased intensive care unit stay time, and major adverse events after transplantation. The multivariate Cox regression analysis suggested that neurological complications (p = 0.001), liver damage (p = 0.011), and respiratory complications (p = 0.044) were independent risk factors for overall mortality after HTx. Combining indicators presented a good predicting effect of peritransplant period mortality (AUC = 0.718).

Conclusion: The mortality of HTx was significantly related to the high-amount transfusion of RBC and plasma. Comprehensively considering the components of blood transfusion obtained better predictive results of peritransplant period survival than solely considering a single component.

Keywords: blood transfusion; heart transplantation; plasma; platelet; red blood cell.

Figure 1

All 568 cardiac transplant patients at Wuhan Union Hospital from January 2015 to December 2020 were recruited. Patients with ages under 18 years, umpteen incomplete follow-up and transfusion data, or receiving retransplant were excluded from the study. Among the 568 recipients, 76 were under the age of 18, 74 had umpteen incomplete follow-up data, and 2 received retransplant. About 416 recipients were finally enrolled in the study cohort according to the inclusion criteria.

Figure 2

Receiver operating curves analysis was performed on the whole population. According to Youden index, the transfusion cut-off values of three blood components were 9 units for red blood cell (AUC = 0.63, 95%CI 0.564, 0.697), 2 packages for platelet (AUC = 0.582, 95% CI 0.514, 0.651), and 1,250 ml for plasma (AUC = 0.566, 95% CI 0.493, 0.639).

Figure 3

The scores of three components presented skewed distribution prior to the propensity score matching, and approximately normally distributed after match. The score distributions were separately shown as below, in which the study population was divided into two groups based on the transfusion amount of: (A) red blood cell (before matching); (B) red blood cell (after matching); (C) platelet (before matching); (D) platelet (after matching); (E) plasma (before matching); (F) plasma (after matching).

Figure 4

Kaplan–Meier analysis estimates survival after heart transplantation with log-rank test, comparing the outcomes between high amount transfusion group and low amount transfusion group. The analysis results were separately shown below, in which the study population was divided into two groups based on the transfusion amount of (A) Red blood cell (beforematching), p < 0.0001; (B) red blood cell (aftermatching), p = 0.0028; (C) platelet (beforematching), p = 0.00017; (D) platelet (aftermatching), p = 0.072; (E) plasma (beforematching), p < 0.0001; (F) plasma (aftermatching), p = 0.012.

Figure 5

Combining the transfusion amount of RBC, platelet, and plasma together had a better accuracy of peritransplant period prognostic assessment to heart transplantation (AUC = 0.718 vs. AUC = 0.682 for RBC, AUC = 0.615 for platelet, and AUC = 0.584 for plasma).