➢ Hemostatic agents are used during total hip and knee arthroplasty to reduce postoperative blood loss and to avoid the adverse events associated with transfusion of allogeneic blood.
➢ Antifibrinolytic agents act by preventing the degradation of fibrin clot, whereas most other hemostatic agents promote clot formation.
➢ Evidence that a hemostatic agent reduces blood loss does not necessarily translate into a reduction in transfusion requirements.
➢ Intravenous tranexamic acid is the only hemostatic agent to consistently reduce blood loss and transfusion requirements in patients undergoing total hip and knee arthroplasty. Furthermore, intra-articular tranexamic acid has been shown to decrease blood loss and transfusion requirements in patients undergoing total knee arthroplasty.
➢ There is inconclusive and insufficient evidence to support the use of many hemostatic agents, emphasizing the need for more high-quality controlled trials.
Total hip and knee arthroplasty are associated with major perioperative bleeding that often necessitates blood transfusion1. As the demand for these successful procedures continues to grow2, the need for limited resources such as allogeneic blood products may increase accordingly. Furthermore, complications associated with the use of allogeneic blood products, including nosocomial infections3-5, transfusion-related lung injury, and circulatory overload6,7, have prompted efforts toward alternative methods of blood conservation. Controlling intraoperative and postoperative bleeding with use of hemostatic agents is one way of avoiding the need for blood transfusion. Several hemostatic agents are available to surgeons, yet not all technologies and products have been evaluated in the context of total hip and knee arthroplasty, and strong evidence regarding the efficacy of some of these agents is still lacking. The choice to use a hemostatic agent likely will depend on the available clinical evidence regarding its efficacy, safety, cost, and ease of administration. The present report briefly discusses the mechanism of action of these agents and provides a critical review of the literature on a variety of hemostatic agents used during total hip and knee arthroplasty. Clinical care recommendations are based on a critical analysis of the best available evidence.
Mechanism of Action
Hemostasis is achieved in injured vessels when exposed extracellular collagen matrix stimulates platelet adhesion and primary platelet plug formation, leading to the subsequent activation of coagulation cascade and fibrin clot formation. Once formed, fibrin clots are dissolved by plasmin, the key enzyme in the fibrinolytic system (Fig. 1). The balance between coagulation and fibrinolysis determines the stability of fibrin clots. Therefore, one can classify the action of hemostatic agents as either antifibrinolytic or procoagulant. Antifibrinolytic agents include tranexamic acid, ε-aminocaproic acid, and aprotinin. All other hemostatic agents contribute to fibrin clot formation or platelet adhesion and can be considered as procoagulants. Numerous hemostatic agents are commercially available and have been used in other surgical specialties8-11. However, the focus of the present review is limited to agents that have been evaluated for use during total hip or knee arthroplasty.
In general, antifibrinolytics inhibit the degradation of the formed fibrin clot by inhibiting plasmin. Bovine aprotinin was the first of these agents to be made available on the market. Aprotinin acts by inhibiting kallikrein, a peptidase that is pivotal to the conversion of plasminogen into plasmin. Furthermore, aprotinin directly inhibits plasmin after its conversion from plasminogen12,13. The use of aprotinin was ultimately suspended by U.S. Food and Drug Administration (FDA) in 2007 because of concerns about increased mortality14,15. The drug is currently unavailable for purchase in the U.S. and therefore will not be discussed further in this review. ε-aminocaproic acid and tranexamic acid are synthetic lysine analogs that act via a reversible interaction with plasminogen and plasmin (Fig. 1). These drugs saturate the lysine binding sites of both plasminogen and plasmin, inhibiting them from binding to fibrin13,16. Antifibrinolytics are also used as adjuvants in some topical fibrin sealant preparations to help stabilize the created clot.
Fibrin Sealants and Platelet Gels
Fibrin sealants accomplish their hemostatic action by bypassing the coagulation cascade to the final step of converting fibrinogen to fibrin (Fig. 1). Generally, these preparations supply a fibrin precursor and thrombin, which are stored as two separate constituents in a dual-syringe kit. These constituents are mixed by connecting the syringes to a single lumen, allowing simultaneous delivery to the surgical wound. Thrombin cleaves fibrinogen, creating fibrin monomers that polymerize into a soluble mesh. Factor XIII that is present in the tissue surface subsequently cross-links with this fibrin mesh to form a stable clot and to achieve hemostasis17. First-generation sealants consisted of human fibrinogen and bovine thrombin plus bovine aprotinin as a stabilizer18. The desire to avoid the use of animal-derived blood products prompted a shift to human thrombin18. Second-generation fibrin sealants replaced bovine thrombin with human thrombin and replaced bovine aprotinin with tranexamic acid or completely removed the antifibrinolytic component. Measures taken to reduce the risk of viral transmission, such as donor screening and viral inactivation, contribute to an increase in cost18. The use of autologous fibrin sealants has been introduced to avoid the risks associated with allogeneic blood products. One such system is a patient-derived fibrin sealant, which utilizes 50 mL of the patient’s blood as a source of fibrinogen and prothrombin. The patient’s blood is mixed with an alkaline buffer solution to lower the pH and to activate endogenous prothrombin.
Platelet gels combine autologous platelet-rich plasma with calcium and thrombin. The presence of platelets is believed to improve the strength of the clot and to provide growth factors to stimulate wound-healing. The source of thrombin varies among different preparations. Commercially available centrifugation systems are used to produce autologous platelet-rich plasma. Platelet-poor plasma is generated as a byproduct of the separation process and can be used as an autologous fibrin sealant when combined with thrombin and calcium19.
Thrombin also can be used alone or mixed with a flowable matrix. One product contains human thrombin mixed with bovine-derived flowable matrix. Other products use a recombinant human thrombin that can be sprayed on bleeding surfaces or mixed with a hemostatic matrix. The use of a hemostatic matrix provides a physical matrix for clotting initiation and platelet adherence. These commercial kits do not include fibrinogen; instead, they require contact with blood from the surgical wound and bleeding surfaces of bone as a source of fibrinogen.
Potential drawbacks to autologous fibrin sealants and platelet gels include cost, the risk of contamination, and the volume of blood needed in autologous preparations. The cost of these agents can reach up to $1400 per kit, and an additional cost of $7495 to $18,500 for a centrifugation system for autologous preparations19. The amount of blood needed to generate the fibrin precursor solution or platelet-rich plasma varies from 10 to 60 mL of whole blood. It is also important to take into account the processing time needed for autologous preparations, ranging from sixteen to thirty-three minutes19. Furthermore, these systems rely on an intact clotting mechanism, so any antiplatelet or anticoagulant medication must be halted in order for the activation process in the wound to take place.
Desmopressin is a synthetic analog of the antidiuretic hormone vasopressin that increases the levels of factor VIII and von Willebrand factor (vWF), enhancing primary hemostasis and platelet adherence. Its use is indicated for patients with clotting disorders such as hemophilia A, von Willebrand disease, or platelet dysfunction12,16. The effect of desmopressin in reducing postoperative bleeding also been has tested in otherwise healthy individuals undergoing total hip and knee arthroplasty20.
Critical Review and Analysis
The studies that were considered for this review included full articles, published in the peer-reviewed literature, evaluating the use of hemostatic agents in patients managed with total hip and knee arthroplasty. An extensive search of the National Institutes of Health PubMed database was performed to identify all epidemiological studies evaluating the use of hemostatic agents in patients managed with total hip and knee arthroplasty up to November 2012. The keywords used in this search included “arthroplasty, replacement, hip,” “arthroplasty, replacement, knee,” and “hemostatic agent” as well as a comprehensive list of hemostatic agents reported in the surgical literature (see Appendix)9,19. All literature searches were supplemented with manual screening of bibliographies of recent systematic reviews and other review articles for potentially relevant citations. Animal and in vitro studies, retrospective cases series, case reports, editorials, commentaries, and studies with fewer than ten patients enrolled were excluded. Studies that did not quantify blood loss or transfusion requirements as primary outcomes also were excluded.
For the critical analysis component of this review, only studies with the highest level of available evidence were included when generating clinical care recommendations. Randomized controlled trials were considered first. In the absence of two or more randomized controlled trials, we sequentially searched for prospective controlled trials, prospective comparative studies, and retrospective comparative studies. Randomized controlled trials were categorized as Level-I studies, but the level of evidence was reduced by one level if the study did not involve adequate randomization, allocation concealment, or patient, caregiver, or evaluator blinding. Grades of recommendation were assigned on the basis of the level of evidence and consistency of findings21,22.
Tranexamic acid is by far the most widely studied hemostatic agent used during primary total knee and hip arthroplasty, with more than thirty-five randomized controlled trials and more than ten meta-analyses having been performed. The efficacy of tranexamic acid in reducing blood loss and transfusion requirements after total hip and knee arthroplasty has been demonstrated invariably in all meta-analyses (Table I)23-28. However, concerns regarding the safety of the systemic administration of tranexamic acid and the risk of thromboembolic events in high-risk patients have limited the widespread adoption of this medication29-31. In their recent meta-analysis, Yang et al. evaluated the effectiveness and safety of intravenous tranexamic acid in a total of 837 patients undergoing total knee arthroplasty who were enrolled in randomized controlled trials28. Their analysis showed that the amount of blood loss per patient was significantly less in the tranexamic acid group as compared with the placebo group, with a weighted mean difference of −505 mL (p < 0.00001). The odds of receiving a blood transfusion also was significantly lower in the tranexamic acid group (odds ratio [OR], 0.16; 95% confidence interval [CI], 0.10 to 0.25; p < 0.00001). Tranexamic acid did not increase the odds of developing deep-vein thrombosis (OR, 0.75; 95% CI, 0.34 to 1.67; p = 0.48) or pulmonary embolism (OR, 0.65; 95% CI, 0.18 to 2.33; p = 0.50).
Sukeik et al.26 performed a meta-analysis of seven studies comprising 350 patients to investigate the effectiveness of tranexamic acid in the setting of total hip arthroplasty. Their analysis revealed a significant reduction in total blood loss by a mean of 289 mL (p < 0.001) and a significant reduction in transfusion requirements by 20% (risk difference, −0.20; 95% CI, −2.9 to −0.11; p < 0.001), without an increase in deep-vein thrombosis, pulmonary embolism, or other complications. Most of the trials that were included in these meta-analyses were powered to assess the efficacy and safety of tranexamic acid in patients managed with elective total hip arthroplasty, with high-risk patients being excluded; therefore, no definite conclusions regarding the safety of tranexamic acid can be made. Nevertheless, a stronger argument for the safety of intravenous tranexamic acid comes from the trauma literature. The CRASH-2 (Clinical Randomization of an Antifibrinolytic in Significant Hemorrhage-2) trial32, a multinational trial in forty countries, included 20,211 adult trauma patients with substantial bleeding who were randomized to receive tranexamic acid or placebo. The results showed that the early administration of tranexamic acid reduced the risk of mortality due to any cause (14.5% compared with 16.0%; p = 0.0035), with no apparent increase in vascular occlusive events such as myocardial infarction or pulmonary embolism (0.3% compared with 0.5%; p = 0.096). Similarly, the Military Application of Tranexamic Acid in Trauma Emergency Resuscitation (MATTERs) study33 retrospectively evaluated the use of tranexamic acid in 896 patients with more severe combat injuries and demonstrated a 6.5% absolute risk reduction in mortality. While the numbers of deep-vein thromboses and pulmonary emboli were too small to assess any risk associated with the use of tranexamic acid, the authors of both studies recommended the incorporation of intravenous tranexamic acid into clinical practice.
There is considerable heterogeneity between clinical trials with regard to administration dose, ranging from low dose (10 to 30 mg/kg) to high dose (150 mg/kg)28. Different methods of delivery include a single intravenous bolus given preoperatively, repeated boluses, prolonged infusion, and intra-articular injection26. Lin et al.34 conducted a randomized controlled trial involving 150 patients undergoing total knee arthroplasty in which a one-dose intravenous injection of 10-mg/kg tranexamic acid (administered five minutes before deflation of the tourniquet) was compared with a two-dose injection (with one dose administered before skin incision and another administered before deflation of the tourniquet). Both regimens reduced blood loss and transfusion requirements, but there were no differences between the one and two-dose groups, suggesting that one intraoperative dose of tranexamic acid is effective for reducing blood loss and transfusion requirements.
The use of intra-articular tranexamic acid has been investigated to avoid the potential risks of thromboembolic events associated with systemic administration35-37. Wong et al.35 used two different dosages of intra-articular tranexamic acid (1.5 and 3 g) in a study of 124 patients undergoing total knee arthroplasty and found a statistically significant reduction in blood loss at both dosages when compared with placebo. However, their study was not powered to detect a difference in transfusion requirements, and only a trend toward fewer transfusions in the treatment groups was reported. Two other, smaller randomized controlled trials36,37 also showed a reduction in blood loss without a reduction in transfusion requirements when intra-articular tranexamic acid was used for patients undergoing total knee arthroplasty. Seo et al.38 conducted a placebo-controlled randomized trial involving 150 patients undergoing total knee arthroplasty and showed that the least amount of blood loss and fewest transfusions occurred in the patients who received intra-articular (transfusion rate, 20%) rather than intravenous tranexamic acid (34%) or placebo (94%). More large randomized controlled trials are needed to evaluate the efficacy of topical application of tranexamic acid and its safety, especially in total hip arthroplasty.
There is a sufficient number of randomized controlled trials supporting the use of intravenous tranexamic acid to reduce blood loss and blood transfusion in association with both total hip and knee arthroplasty (Grade of Recommendation, A) (Table II). The intra-articular administration of tranexamic acid showed promise for reducing blood loss and transfusion requirements in patients managed with total knee arthroplasty (Grade of Recommendation, B). We are not aware of any studies that have evaluated the effects of tranexamic acid in patients undergoing revision hip and knee arthroplasty.
The search revealed only three randomized clinical trials on the use of ε-aminocaproic acid in patients managed with primary total hip or knee arthroplasty (Table I)39-41. Two studies involving forty-five and forty-nine patients who were managed with total hip arthroplasty demonstrated the efficacy of ε-aminocaproic acid for reducing mean blood loss in comparison with placebo39,40, although neither study showed a difference in transfusion requirements between treatment arms. Camarasa et al.41 compared tranexamic acid and ε-aminocaproic acid in a double-blind, placebo-controlled trial involving 128 patients who were managed with total knee arthroplasty. Patients receiving any antifibrinolytic agent (either tranexamic acid or ε-aminocaproic acid) had significantly lower blood loss compared with those who received placebo (1099 compared with 1784 mL; p < 0.001). Furthermore, allogeneic blood transfusion was performed for 7.5% of patients in the antifibrinolytic group, compared with 38.3% in the placebo group (p < 0.001). The authors did not report a separate comparison between ε-aminocaproic acid and placebo. An increased transfusion rate was noted in the ε-aminocaproic acid group compared with the tranexamic acid group, although the difference was not significant (12.5% compared with 2.8%; p = 0.185). The number of patients studied could have impeded the detection of significant differences found between the two antifibrinolytic groups.
Although there is good evidence to support the efficacy of ε-aminocaproic acid in reducing blood loss after total hip and knee arthroplasty, the evidence is less compelling with regard to transfusion requirements (Grade of Recommendation, A).
Fibrin Sealants and Platelet Gels
A variety of fibrin sealants and platelet gels have been tested for primary total hip and knee arthroplasty, whereas none have been tested for revision procedures (Table I). It should be noted that the surgeons in these trials were not blinded to the treatment as no placebo was used. Allocation concealment was still possible as some investigators revealed the treatment allocation right before capsule closure to prevent any bias or deviation from standard hemostatic techniques. Several randomized controlled trials42-46 evaluated the use of human-derived fibrin sealant in total knee arthroplasty. Levy et al.42 reported the results of what we believe to be the first multicenter randomized controlled trial, which involved fifty-eight patients who were managed with total knee arthroplasty. Postoperative blood loss, mean decrease in hemoglobin concentration, and calculated blood loss were all lower in the fibrin sealant group as compared with the standard treatment group. Transfusion requirements were also significantly lower in the fibrin sealant group than in the standard treatment group (20% compared with standard 83%; p = 0.004). Other studies showed inconsistent results with regard to transfusion requirements (Table I) and lacked proper randomization or allocation concealment43-47. Limited evidence exists on other commercially available human-derived fibrin sealants48,49, human thrombin50,51, and bovine-derived thrombin52. Skovgaard et al.48 performed a randomized controlled trial evaluating a new-generation fibrin sealant in a study of twenty-four patients undergoing bilateral total knee arthroplasty. The left knee was operated on and randomized first (treatment or placebo), and the contralateral knee was assigned to the opposite study arm. Drain output was the main outcome and did not show a difference between the two groups, whereas transfusion requirements could not be evaluated because of the bilateral setting. Only one randomized controlled trial evaluated the use of human thrombin combined with bovine gelatin matrix in total knee arthroplasty50. A total of 196 patients were enrolled, but no difference in total drain output or transfusion requirement was observed.
In patients undergoing total hip arthroplasty, fibrin sealants may have an effect on reducing blood loss, but no studies have demonstrated a reduction in transfusion requirements, to our knowledge. The use of fibrin sealants during total hip arthroplasty also showed a decrease in blood loss, but no reduction in transfusion requirement was demonstrated. Wang et al.53 conducted a single-blind randomized multicenter trial of eighty-one patients to examine the role of fibrin sealant in total hip arthroplasty. The results of their intention-to-treat analysis showed that the mean blood loss in the fibrin sealant group was 23.5% less than that in the control group (626 compared with 819 mL; p = 0.014), but the difference between the two groups in terms of the total number of transfused units was not significant. In another open-label randomized trial, Lassen et al.54 compared the autologous fibrin sealant system with the standard (no hemostatic treatment) in a study of eighty patients undergoing total hip arthroplasty. The use of autologous fibrin sealant system did not reduce blood loss by a significant amount as determined on the basis of drainage volumes, transfusion requirements, and changes in hematocrit and hemoglobin levels. The reduction of blood loss obtained with autologous fibrin sealant was of a similar magnitude to the volume donated by the patient for preparation of the fibrin sealant. Mawatari et al.55 also used an autologous fibrin sealant in a randomized trial of 100 patients undergoing total hip arthroplasty and found a reduction in blood loss but no difference in terms of transfusion requirements.
Three randomized trials56-58 compared fibrin sealants with tranexamic acid and standard treatment in patients undergoing total hip and knee arthroplasty. In all three trials, the fibrin sealant reduced blood loss in comparison with standard treatment but did not reduce the need for transfusion. Furthermore, there was no significant difference between fibrin sealant and tranexamic acid with regard to blood loss or transfusion requirements. In all of the trials evaluating fibrin sealants prepared from allogeneic plasma, there were no cases of viral transmission and the incidence of moderate to severe adverse events was comparable between the treatment and control groups. Overall, most studies have shown a reduction in blood loss in association with the use of human fibrinogen and thrombin, but the evidence regarding its efficacy in reducing the need for transfusion is inconsistent (Grade of Recommendation, C). There is insufficient evidence to recommend the use of other human-derived or bovine-derived preparations (Grade of Recommendation, I).
Few studies have been published on the use of autologous platelet gels in patients undergoing primary total knee arthroplasty. Horstmann et al.59 performed a randomized controlled trial of forty patients to evaluate the use of autologous platelet gel in total knee arthroplasty. Platelet-rich plasma was prepared by means of centrifugation. There was no significant difference between the platelet gel group and the control group in terms of either the decrease in hemoglobin postoperatively or the transfusion requirements, although this result was likely due to the study being underpowered. Other retrospective studies on the use of autologous platelet gel (platelet-rich plasma) have shown conflicting results60-62. In a prospective observational study of 165 patients who were managed with total knee arthroplasty, Everts et al.63 demonstrated a reduction in transfusion requirements when autologous fibrin sealant was combined with platelet gel. The results of that study were limited by the fact that drains, which may increase perioperative blood loss64, were used in the control group but not in the treatment group. Our group conducted a randomized controlled trial of 100 patients undergoing primary total knee arthroplasty to evaluate the efficacy of a particular platelet gel in reducing transfusion requirements65. No drains were used, and aspirin was used for thromboprophylaxis. Transfusion requirements, as determined by a blinded investigator using standardized criteria, were significantly lower in the treatment group (no blood transfusions) as compared with the control group (five transfusions) (p = 0.007). Overall, there is insufficient evidence to recommend for or against the use of platelet gels, and more randomized controlled trials are needed (Table II).
Six randomized, placebo-controlled studies addressing the use of desmopressin in patients undergoing primary total hip or knee arthroplasty were identified and reviewed66-71. Two studies involving fifty66 and eighty68 patients undergoing total hip arthroplasty examined the effect of using two 0.3-µg/kg doses of dextran (total dose, 0.6 µg/kg) for thromboprophylaxis and as a plasma substitute. Dextran has antithrombotic effects via the dilution of coagulation factors, decreased platelet aggregation and adhesion, decreased vWF levels, and enhanced fibrinolysis72,73. The investigators administered desmopressin in order to reverse the effects of dextran and thereby reduce blood loss and transfusion requirements. Although desmopressin reduced total calculated blood loss in one study, neither study demonstrated a difference in terms of transfusion requirements as compared with placebo. Two other studies67,69 investigated the administration of desmopressin without the use of dextran as a resuscitative fluid. Karnezis et al.67 used a single 0.3-µg/kg dose in ninety-two patients who were managed with primary total hip or knee arthroplasty. In that study, the use of desmopressin did not demonstrate a benefit in terms of reducing blood loss or transfusion requirements and did not affect bleeding time. Leino et al.69 compared high-dose (0.4-ug/kg) and low-dose (0.2-µg/kg) desmopressin with placebo (50 mL of saline solution) in a study of seventy-five patients with rheumatoid arthritis. Total blood loss, estimated from suction bottles and weighing swabs, was not significantly different between the three groups (p = 0.50). However, fewer transfusions were needed in the high-dose desmopressin group (mean, 3.6 units) as compared with the low-dose desmopressin group (mean, 4.4 units) (p = 0.009) and placebo groups (mean, 4.5 units) (p = 0.011). Two randomized clinical trials compared desmopressin with tranexamic acid in patients managed with total knee arthroplasty70,71. The dose of desmopressin was 0.3 µg/kg. In both studies, patients who received desmopressin had higher allogeneic blood transfusion rates and lower postoperative hematocrit compared with those who received tranexamic acid. A meta-analysis that included 2488 patients evaluated the incidence of adverse events related to the use of desmopressin in both cardiac and noncardiac procedures did not demonstrate a significant increase in thromboembolic events20. The majority of evidence suggests that desmopressin is ineffective for reducing blood loss or transfusion requirements after total knee or hip arthroplasty (Grade of Recommendation, A).
The available literature examining hemostatic agents shows substantial heterogeneity in terms of study design. Variations in factors such as drain usage, postoperative thromboprophylaxis regimen, treatment dose, method of administration, transfusion protocol, and blood loss measurements make the interpretation of such studies difficult. However, tranexamic acid consistently has been shown to be effective for reducing blood loss and transfusion requirements, regardless of how one defines blood loss or transfusion triggers (Grade of Recommendation, A). A single (or repeated) inexpensive intravenous injection of tranexamic acid is more practical and cost-effective74-78 than other biological agents. Although fibrin sealants and platelet gels have been shown to reduce blood loss, the evidence regarding their effect on blood transfusion is inconclusive or insufficient. The process of obtaining and processing autologous blood in some of these preparations makes them less cost-effective. Furthermore, the efficacy of these agents may be related to their method of application and to surgeon experience79. Other hemostatic agents do not have sufficient evidence to recommend for or against their use in total hip and knee arthroplasty (Grade of Recommendation, I). We did not find any randomized controlled trials evaluating the use of hemostatic agents in revision hip and knee arthroplasty. Our group recently completed a randomized controlled trial evaluating a platelet gel preparation for use in primary total hip arthroplasty (submitted for publication), and a randomized trial investigating the use of the same platelet gel in revision total knee arthroplasty is currently underway.
A table showing the PubMed search strategy is in Appendix I.
Investigation performed at the Department of Orthopaedic Surgery, Cleveland Clinic, Cleveland, Ohio
Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. One or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.
- Copyright © 2014 by The Journal of Bone and Joint Surgery, Incorporated