Intravenous Tranexamic Acid Reduces Post-Operative Bleeding and Blood Transfusion in Patients Undergoing Aortic Surgery: A PRISMA-Compliant Systematic Review and Meta-Analysis

Background: Tranexamic acid (TXA), an antifibrinolytic agent, has been demonstrated to reduce blood loss and transfusion requirements in both cardiac and non-cardiac surgery. However, the evidence regarding the efficacy of intravenous TXA in aortic surgery has been seldomly analyzed. Therefore, the current study was performed to address this question. Methods: Searches of PubMed, EMBASE, OVID, Cochrane Library and CNKI were conducted comprehensively for randomized controlled trials (RCTs) comparing intravenous TXA versus no-TXA. Independently and in duplicate, we reviewed titles, abstracts and full-text articles, extracted data and evaluated bias risks. A random effect or fixed effect model was utilized to pool data. Results: The database search yielded 4 RCTs involving 273 patients. Meta-analysis revealed that, there was a significant reduction in bleeding volume within the first 4 hours post-operatively [(weighted mean difference (WMD) = –74.33; 95% confidence interval (CI): –133.55 to –15.11; p = 0.01)], and the first 24 hours post-operatively [(WMD = –228.91; 95% CI: –352.60 to –105.23; p = 0.0003)], post-operative red blood cell (RBC) transfusion volume [(WMD = –420.00; 95% CI: –523.86 to –316.14; p < 0.00001)], fresh frozen plasma (FFP) transfusion volume [(WMD = –360.35; 95% CI: –394.80 to –325.89; p < 0.00001)] and platelet concentrate (PC) transfusion volume [(WMD = –1.27; 95% CI: –1.47 to –1.07; p < 0.0001)] following intravenous TXA administration. In addition, intravenous TXA administration significantly decreased the incidence of postoperative complications (53/451 (8.2%) vs. 75/421 (13.9%); odds ratio (OR) = 0.47; 95% CI: 0.30 to 0.75; p = 0.001), according to this present meta-analysis. Conclusions: The current study preliminarily demonstrated that, TXA significantly reduced postoperative bleeding, blood transfusion requirements and postoperative complications among patients undergoing aortic surgery. More well-designed studies are warrant to confirm the efficacy and safety of intravenous TXA in patients undergoing aortic surgery.


Introduction
Aortic aneurysm surgery was first performed by Cooley in 1952 [1].This type of surgery is commonly accompanied by excessive bleeding after surgery, which often requires extensive blood transfusion.Numerous factors contribute to this problem.Activation of the fibrinolytic system secondary to rapidly formed thrombi, extracorporeal circulation, deep hypothermia and circulatory arrest (DHCA) [2,3] have been certified as an important mechanism for surgery-related bleeding.And all these factors complicated post-operative patient outcomes.
The efficacy of TXA in aortic surgical patients remains unknown.Therefore, the current study was performed to address this question.

Inclusion Criteria
Our study included all randomized controlled trials (RCTs) comparing TXA with no treatment or placebo in patients undergoing open aortic surgery in terms of efficiency and safety.Tranexamic acid, no treatment and placebo groups were eligible only from studies with other comparator drugs.The primary outcomes were the intra-and post-operative blood loss, allogeneic transfusion and incidence of reoperation due to post-operative bleeding.The secondary outcomes were the incidences of myocardial infarction (A new Q wave on the electrocardiogram, a creatine kinase MB/creatine kinase ratio greater than 10%, and a troponin I value of more than 0.1 ng/dL on the troponin I test) [6], stroke, acute lung injury (hypoxemia occurred within 72 h after surgery, an oxygenation index (arterial oxygen partial pressure/inhaled oxygen concentration) <150 mmHg) [13], pulmonary infection, acute renal insufficiency (two times the baseline creatinine level or dialysis needed) [6], gastrointestinal bleeding, hepatic insufficiency (glutamate-pyruvate transaminase (GPT) >200 U/L, total bilirubin (TBIL) >50 mm/L, and lactate dehydrogenase (LDH) >50 U/L occurring within one week after surgery, with or without clinical manifestations of liver insufficiency) [14].

Exclusion Criteria
Exclusion criteria included: (1) articles published in case reports, abstracts, or reviews; (2) observation or retrospective studies (3) articles without information on outcomes of interest; (4) cell or animal studies; (5) duplicate publications.Titles and abstracts of all the identified eligible studies were independently reviewed to exclude evidently ineligible ones by two authors (BZ and LXH).We further determined whether eligibility for these residual studies was ultimately included by reading the full article.

Study Quality Assessment
The risk of bias was assessed independently by two authors (BZ and LXH), using the tool referred to in the Cochrane Handbook for Systematic Reviews of Interventions [15].Further, two authors (LXH and YTY) independently evaluated the methodological quality of each trial using the modified 7-point Jadad score [16].The trials with 1-3 points were evaluated as low quality, and the trials with 4-7 points were evaluated as high quality.

Data Abstraction
The following data were independently extracted by two authors (BZ and LXH) from the included studies: (1) author, countries and publication years of studies included; (2) total number of participants, age, sex, body weights (BW)/body mass indexes (BMI) of patients in TXA and placebo/blank groups; (3) surgical type; (4) data related to the outcomes of interest in two groups.Data processing was completed with all disagreements resolved through discussion among all authors.

Statistical Analysis
The datum was analyzed using RevMan 5.3 (Cochrane Collaboration, Oxford, UK).For dichotomous data, we used a pooled odds ratio (OR) and 95% confidence interval (CI), while for continuous data, we used weighted mean difference (WMD) and 95% CI.Depending on whether significant heterogeneity (I 2 >50%) existed for each outcome, randomized-effects or fixed-effects models were utilized.We examined how statistical models affected estimates of management effects in sensitivity analyses.All analyses adopted the fixed-effects model were twice assessed via randomized-effects model and vice versa.Furthermore, sensitivity analyses were conducted to exam the effect of individual studies on the overall results.An investigation of the publication bias was conducted using funnel plots of the outcomes.In all cases, the p-value was two-sided, and p < 0.05 was considered statistically significant.

Search Results
Fig. 1 illustrates how database searches identified seven articles for full evaluation.There were four studies eligible for inclusion in the meta-analysis [6,8,10,11].Table 1 (Ref.[6,8,10,11]) presents a description of these articles.Of the 4 references, two [6,8] were written in English (2 from Italy), and the other two [10,11] were in Chinese.

Included Trials Characteristics
As showed in Table 1, among the four trials, three included patients undergoing thoracic aortic aneurysm surgery [6,10,11], one included patients with surgery of abdominal aortic aneurysm [8].There were 273 patients enrolled in the 4 eligible trials, and 139 were allocated to the TXA group and 134 to the Control (placebo) group.

Sensitivity Analyses and Publication Bias
According to sensitivity analysis, statistical model choice did not affect treatment effects.In addition, sensitivity tests were conducted to determine how individual studies influence overall effects by removal of each study.For high heterogeneity outcomes, we also performed sensitivity tests to examine the influences of the overall treatment effects on it by exclusion of some studies, whereas we did not find any contradictory results.A funnel plot analysis of primary and secondary outcomes did not reveal any significant publication bias.

Discussion
As far as we know, this present meta-analysis is the initial study to analysis the effects of intravenous TXA on perioperative blood loss, perioperative blood transfusion  Aortic surgery involves high levels of bleeding and perioperative blood transfusions as a result, which frequently complicates patient outcomes.The preoperative factors include the diseased aortic fragile tissues due to the formation of a false lumen and increased fibrinolysis caused by the tissue factors and then activate coagulation factor vii to start the extrinsic coagulation pathway [17].Several intraoperative factors are also involved, as well as a large surgical surface, the application of extracorporeal circulation, and DHCA [18,19].All these factors play parts in the complex changes of hemostatic and the consequent abnormal perioperative blood loss frequently occurred in patients undergoing aortic surgery.TXA is an antifibrinolytic agent which is widely used in the management of reduction of intra-and postoperative bleeding and transfusion requirements.It has been certified that TXA can reduce blood loss and transfusion requirements effectively in cardiac [20] and non-cardiac surgical patients [21,22].Similar results were obtained by Ahn et al. [5] in a study which retrospectively studied data from 55 adult patients who underwent acute aortic dissection surgery between April 2008 and April 2010.Moreover, Makhija et al. [7] also reported TXA had equal effect in reducing the perioperative bleeding and transfusion requirements in patients undergoing surgery for thoracic aortic compared with epsilonamino-caproic acid.Similar findings [6,8,10,11] supported the conception that the intravenous injection of TXA played a significant role in the reduction of postoperative bleeding and transfusion requirement after aortic surgery.
Our meta-analysis also showed that no significant difference was detected in the rate of FFP and platelets postoperatively transfusion among the groups.However, in the rate of reoperation for bleeding and postoperative transfusion of RBC, TXA group showed significantly less rate compared to placebo.Casati et al. [6], Xu et al. [10] and Zhang et al. [11] demonstrated similar results in their respective studies comparing the groups.Decreasing operative bleeding would lessen the risks and costs related to blood transfusion.In addition, a higher rate of surgical reexploration occurred in patients with excessive bleeding [23].Contrarily, some trials [5][6][7][8] reported that TXA administration had no effect on reducing the rate of reoperation for bleeding or the rate of postoperative transfusion of RBC in aortic surgery.It is assumed that a much larger sample size would be necessary to detect potential differences between groups of related outcomes.Hopefully further well-designed trials will address this issue in a more comprehensive way.
The current study demonstrated that, intravenous TXA was associated with lower incidences of postoperative complications [6,8,10,11].Furthermore, no differences between the groups were observed in terms of thromboembolic complications (myocardial infarction or stroke).As we known, it is associated with poorer postoperative outcomes when excessive bleeding leads to blood transfusions.In fact, it has been proven in two large observational studies that perioperative transfusions are associated with increased rates of morbidity and mortality during the 30 days following major vascular surgery [24,25].In addition, an allogenic blood supply is limited and should only be used when absolutely necessary.Thus, Strategies or drugs that can reduce bleeding and transfusion requirements may improve clinical outcomes.However, the use of tranexamic acid was still controversial, because sporadic reports of thrombotic events and stroke was reported.Recently, an analysis of over 5000 cardiac surgery patients showed that tranexamic acid administration was not related with a higher risk of thrombotic complications [26], which was consistent with the results of the present study.However, Zhou et al. [27] reached the opposite conclusion in a retrospective study that the use of intraoperative TXA during cardiac surgery was associated with postoperative strokes.
Several limitations are present in this study.It cannot be denied that, by pooling a large number of small, low-quality studies, the meta-analysis is able to expand the power of the analysis, but its limitations are also evident.These include multiple TXA dosages, limited trial scale and quality and heterogeneity issues of included studies.We aimed to provide some evidence about this clinical controversy with our meta-analysis.The only way to prove this is through more well-designed, large-scale randomized trials.

Conclusions
The present study provided some preliminary evidence that intravenous administration of TXA in aortic surgical patients was effective not only in reducing blood loss and transfusion requirement, but also so in lowering the incidences of postoperative complications.