Previous meta-analyses established that tranexamic acid confers benefits when used during total hip arthroplasty (THA). However, 2 of these meta-analyses included a variety of routes of administration of tranexamic acid in THA (topical, intravenous, oral, and intra-articular), another meta-analysis included a variety of antifibrinolytic drugs (not restricted to a single drug), and the final meta-analysis included nonrandomized controlled trials. This meta-analysis focused on a single medication, tranexamic acid, administered in a specific way, intravenously in patients undergoing primary THA, using data reported only in randomized controlled studies. Outcomes were restricted to blood loss, allogeneic transfusion rates, and complications. Other outcomes, such as return to function or clinical scores, could not be evaluated because of lack of consistent reporting. To better understand the effects of intravenous tranexamic acid in THA on clinical outcomes, such as recovery, return to function, and patient-reported outcome measures, it would be helpful to have more controlled trials examining these measures in a standardized manner. Intravenous tranexamic acid was beneficial for blood loss intraoperatively, blood loss through drains, and total blood loss during hospitalization, in addition to reducing allogeneic transfusion rates. No difference between intravenous tranexamic acid and placebo was found for most complications, except deep venous thrombosis, which showed favorable results with placebo. [Orthopedics.2016; 39(5):e883–e892.]
Tranexamic acid is a synthetic amino acid and competitive inhibitor of plasminogen that acts to decrease fibrinolysis and thus decrease bleeding and blood loss as a result of the trauma of surgery and the resulting release of tissue plasminogen activator.1,2 The body naturally inhibits fibrinolysis by 24 hours after surgery, but antifibrinolytics such as tranexamic acid may block the activation of plasminogen to plasmin earlier and therefore decrease perioperative blood loss.2–4 Tranexamic acid has been used successfully to decrease blood loss in numerous surgical specialties, especially cardiac surgery.2,5,6 More recently, tranexamic acid therapy was used in total joint arthroplasty, with a similar reduction in allogeneic transfusion rates. Patients undergoing primary total hip arthroplasty (THA) often have perioperative blood loss and require blood transfusions.7,8 This blood loss is associated with longer hospital stays and delayed rehabilitation and may be poorly tolerated in patients with comorbidities.8 The primary effect of tranexamic acid on patient care and outcomes in the perioperative period is a reduced need for allogeneic transfusion immediately after THA.9,10 Additionally, tranexamic acid eliminates or reduces several risks and complications of blood transfusion, including transfusion reactions, infections, fluid overload, and altered mental status, all of which may lead to prolonged hospitalization and convalescence and additional morbidity.10–13
Review of the literature identified 4 meta-analyses of tranexamic acid in primary THA that were published in the past 5 years; the focus of these meta-analyses was not tightly bound. Of these studies, 2 examined tranexamic acid in THA and total knee arthroplasty (TKA) and 2 focused on tranexamic acid in THA only.6,14–16 Each of these studies measured allogeneic transfusion rates, blood loss, and various complications, especially thromboembolic events. Alshryda et al14 also included functional outcome measures and quality of life. Their study focused on topical application of tranexamic acid in THA and TKA and found that it was a safe and effective method for reducing the need for blood transfusions. However, 6 months after total joint arthroplasty, no difference was found in functional outcomes or quality of life. Gandhi et al15 did not focus on a particular route of administration of tranexamic acid in either THA or TKA, examining intravenous, oral, intra-articular, and topical applications equally and comparing them with placebo or no treatment. Their meta-analysis showed that tranexamic acid resulted in less total blood loss and fewer allogeneic blood transfusions, with no increase in deep venous thrombosis. Sukeik et al6 conducted a meta-analysis not by comparing tranexamic acid with placebo in primary THA but by comparing tranexamic acid with placebo, another antifibrinolytic agent, or no treatment. As with the other meta-analyses, these authors found a significant reduction in allogeneic blood transfusion rates, with no significant difference in rates of deep venous thrombosis, pulmonary embolism, infection, or other complications. The fourth recent meta-analysis and systematic review did not restrict the pool of data to randomized controlled trials, limiting its usefulness.16 Wang et al16 focused on topical application of tranexamic acid in primary THA; the primary outcomes were blood loss and transfusion rates. This study found significant differences in total blood loss, postoperative blood loss through drainage, and transfusion rates, but no significant differences in thromboembolic events or wound infections.
The current study focused on intravenous tranexamic acid vs placebo in primary THA, with primary outcomes of blood loss, proportion of patients receiving allogeneic blood transfusions, and complication rates. The data pool for the current study included 3 new randomized controlled trials that were not included in previous meta-analyses.17–19
The methods and reporting of the review were in accordance with the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) Statement.20 All included studies were randomized controlled trials that met the eligibility criteria: tranexamic acid administered intravenously in patients undergoing primary THA.
A comprehensive search of the PubMed bibliographic database was performed to retrieve articles published from January 2000 to November 2015. The search for pertinent studies was conducted with the following terms: “total hip” or “revision hip” and “replacement” or “arthroplasty” and “antifibrinolytic” or “tranexamic acid” or “TXA” or “Lysteda” or “aprotinin” or “Traysol” or “aminocaproic acid” or “Amicar.” Search parameters were restricted to English language studies of adult populations. Additional searches were conducted from the reference lists of meta-analyses published on tranexamic acid in total joint arthroplasty.6,14–16,21 After further screening of the search results, only studies in which tranexamic acid was administered intravenously in primary THA were used. In-sufficient data were available to consider other modes of delivery, and inclusion of other antifibrinolytics was considered to make the pooling of data unclear. Studies included in the qualitative analysis were peer-reviewed, primary data articles that reported THA performed with either tranexamic acid or placebo. Gray literature was not included.
The search yielded 134 references (Figure 1). Titles and journals of origin were screened by 1 of the authors (S.G.C.), and records were excluded based on 12 criteria: not THA, not antifibrinolytic, not clinical, not comparative, not tranexamic acid vs placebo, aminocaproic acid, anticoagulant, antifibrinolytic combined with a mechanical process, aprotinin, focus on deep venous thrombosis, data not recoverable (as a result of poor reporting), and use of special anesthesia to control blood loss (n=66). An additional 42 references were excluded according to the same criteria after examination of the full-text articles. Potentially eligible articles (n=26) were further assessed by review of abstracts in greater detail (S.G.C.). Included randomized controlled studies had to satisfy a single criterion: they had to compare intravenously administered tranexamic acid with a placebo administered intravenously during primary THA.
 |
Figure 1: Flow chart showing the number of articles selected from PubMed searches and other sources and the number remaining after exclusion criteria were applied. Abbreviations: RCT, randomized controlled trial; THA, total hip arthroplasty; TXA, tranexamic acid. |
Data were extracted from each report and entered into a standardized data extraction spreadsheet developed internally by 1 of the authors (S.G.C.) according to principles from the Cochrane Handbook for Systematic Reviews of Interventions.22 Both authors independently extracted data elements from the included studies. Data points entered into the data extraction spreadsheet included the following: country where the study was performed, dosage of tranexamic acid, Centre for Evidence-Based Medicine level of evidence,23 routine procedures related to blood loss, modified Jadad score,24 patient demographics, operative time, intraoperative blood loss, postoperative blood loss through surgical drains, total blood loss, allogeneic blood transfusion rate, and complications related to blood loss and administration of antifibrinolytic agents. For any of these variables to be considered for data analysis, it had to be reported in at least 4 references because the use of fewer than 4 references creates excessive opportunity for bias (eg, to perform data analysis on variable X, 4 references must report variable X for intravenous tranexamic acid and placebo in primary THA). Because data are not always reported consistently or frequently enough to meet this criterion, a limited pool of variables were available for comparison. Patient characteristics, such as age, body weight, height, and percentage of male patients, were available for analysis. Surgical and hospital variables, such as length of surgery, intraoperative blood loss, postoperative blood loss from drains, total blood loss, and allogeneic blood transfusion rate (percentage of patients undergoing THA and receiving a transfusion), were sufficient for analysis. Complication rate data exist as events and nonevents. Therefore, when a study reports zero pulmonary embolisms, for example, it is not included in meta-analysis calculation. For this reason, complication rates were reported by systematic review and data pooling, using only complications reported in at least 4 references.
A total of 16 sources provided data from primary THA comparing 19 intravenous tranexamic acid data sets with 16 placebo data sets, and 993 THA procedures were analyzed (Table 1). All studies were randomized controlled trials.1,3,17–19,25–35 Included studies were from around the globe, predominantly Europe (47% of THA procedures) and Asia (42% of THA procedures). Average modified Jadad score for quality was 6.21 of 8 possible points (Table 1). The remaining 10 articles contributed background information.6,13–16,36–39
   |
Table 1: Characteristics of Oxford Center for Evidence-Based Medicine Level I Randomized Controlled Trials |
Statistical analysis was performed with JMP version 11.1.1 software (SAS Institute Inc, Cary, North Carolina), and Comprehensive Meta-Analysis version 3 software (Biostat, Englewood, New Jersey). Statistical significance was defined as P<.05. Clinical importance is not easily defined by a numerical value and must be determined based on clinical experience. Findings were reported with consideration of statistical significance and the authors' judgment of clinical importance. The authors used JMP Statistical Analysis software to create distributions of various data characteristics and weighted means where possible by using the number of THA procedures of enrolled patients as the denominator.
A random-effects meta-analysis, using Comprehensive Meta-Analysis software, was used to reflect variation in studies. Heterogeneity was tested with Cochran's Q-test, expressed as P and I2, and it was determined that the fixed-effects modality was inappropriate (larger I2 indicates increasing heterogeneity and thus the need for random-effects modeling). According to Borenstein et al,40 random-effects meta-analysis should be used when data are accumulated from a series of studies performed by independently functioning researchers. This form of meta-analysis allows for inclusion of studies, regardless of population size, without assigning too little or too much weight to the data.40 Continuous variables were analyzed with reported means and standard deviations for effect sizes; meta-analysis results were shown as standard difference in means. Meta-analysis results for variables that were reported as incidents or rates were shown as relative risk values.
No differences were found in patient characteristics, indicating a demographically well-matched analysis (Table 2). No statistically significant difference was found for surgical time in the meta-analysis of available data (P=.296) for placebo (weighted mean±SD, 105.08±17.61; n=281) and intravenous tranexamic acid (weighted mean±SD, 117.98±17.74; n=279).
 |
Table 2: Patient Characteristics |
Meta-analysis found significant differences in measures of blood loss, such as intraoperative blood loss, blood loss through drains during immediate postoperative hospitalization, and total blood loss (Figures 2–4 and Table 3). In each instance, intravenous tranexamic acid showed better results than placebo. Lower levels of blood loss occurred with intravenous tranexamic acid.
 |
Figure 2: Intraoperative blood loss (milliliters), shown by a forest plot of pooled standard differences in means. Abbreviations: CI, confidence interval; IV-TXA, treatment group receiving intravenous tranexamic acid; placebo, control group receiving an equal dose of sterile saline; Std diff, standard difference. |
 |
Figure 3: Postoperative blood loss through suction drains (millimeters), shown by a forest plot of pooled standard differences in means. Abbreviations: CI, confidence interval; IV-TXA, treatment group receiving intravenous tranexamic acid; placebo, control group receiving an equal dose of sterile saline; Std diff, standard difference. |
 |
Figure 4: Total blood loss throughout hospitalization for total hip arthroplasty (milliliters), shown by a forest plot of pooled standard differences in means. Abbreviations: CI, confidence interval; IV-TXA, treatment group receiving intravenous tranexamic acid; placebo, control group receiving an equal dose of sterile saline; Std diff, standard difference. |
 |
Table 3: Variables With a Statistically Significant Difference Found From Meta-analysis |
Meta-analysis also found significant differences in the proportion of patients receiving allogeneic blood transfusions (Figure 5 and Table 3), and again intravenous tranexamic acid showed better results than placebo. Lower proportions of patients treated with intravenous tranexamic acid received blood transfusions.
 |
Figure 5: Rates of allogeneic blood transfusion shown by a forest plot of pooled standard differences in means. Abbreviations: CI, confidence interval; IV-TXA, treatment group receiving intravenous tranexamic acid; placebo, control group receiving an equal dose of sterile saline. |
Complication rates were not typically significantly different (Table 4). Only the rate of deep venous thrombosis was significantly different, with patients receiving intravenous tranexamic acid having a higher rate than those receiving placebo. The other complications, pulmonary embolism, clinical wound hematoma, and wound difficulty (ie, difficulty healing and/or superficial infections), were not significantly different between the 2 groups.
 |
Table 4: Variables Analyzed by Systematic Review and Data Pooling |
The cost of perioperative transfusion and blood loss is measured not only financially but also in terms of other resources and outcomes. Patients who have less blood loss and thus do not need allogeneic transfusions avoid some costly side effects.10–13 Gillette et al13 found that although there is an increase in pharmacy costs associated with the use of tranexamic acid, this increase is more than offset by reductions in other costs associated with THA and TKA. Harris et al10 found reductions in facility costs ($286.90/THA procedure before tranexamic acid vs $132.41/THA procedure with tranexamic acid) as well as reductions in man-hours needed to provide care associated with blood transfusions (0.45 man-hours/THA procedure before tranexamic acid vs 0.14 man-hours/THA procedure with tranexamic acid). Slover and Bosco38 examined the cost profile of preoperative tranexamic acid in THA and TKA with regard to major bleeding complications and found that the cost-effectiveness of tranexamic acid is based on baseline blood transfusion rates at surgical centers. In other words, if the baseline transfusion rate is less than 25%, then there is no cost savings.
Although it is important to examine the economic effect of tranexamic acid in THA, it is also important to examine the effect of tranexamic acid in the perioperative and postoperative periods. The current meta-analysis showed that intravenous use of tranexamic acid during THA has positive effects on blood loss and transfusion rates and no negative effect on thromboembolic complications.
There are limitations to the current analysis. First, because this study is a pooled analysis of published results, there is the inherent limitation in how results are reported. This is a common problem with most evidence-based medicine, and for this reason, meta-analyses should look at questions that are defined as specifically as the existing data will allow. Second, the studies used varying routes of administration and dosages of tranexamic acid, and although this study considered only intravenous boluses, the timing and dosage of the bolus varied across the data set.
Although it would be advantageous to discuss other possible effects of intravenous tranexamic acid in THA, such as faster recovery and return to function and improved patient well-being, at this point, it is sufficient to know that tranexamic acid in THA significantly decreases blood loss and transfusion rates without increasing thromboembolic complications. To better understand the effects of intravenous tranexamic acid in THA on clinical outcomes, such as recovery, return to function, and patient-reported outcome measures, it would be helpful to have more controlled trials examining these measures in a standardized manner.
Intravenous tranexamic acid reduces blood loss and allogeneic transfusion rates. This study found no difference between intravenous tranexamic acid and placebo for most complications, except deep venous thrombosis, which showed better results with placebo.
C/ San Pedro de Mezonzo nº 39-41
15701 – Santiago de Compostela
Teléfono: +34 986 417 374
Email: secretaria@sogacot.org
Coordinador del Portal y Responsable de Contenidos: Dr. Alejandro González- Carreró Sixto
