➢ Clinical practice guidelines from the American Academy of Orthopaedic Surgeons (AAOS) and the American College of Chest Physicians (ACCP) and a clinical effectiveness literature review from the Agency for Healthcare Research and Quality (AHRQ) are in agreement about the need for active intervention in preventing venous thromboembolism after total hip arthroplasty or total knee arthroplasty, but a paucity of clinical end-point data provide no basis for the endorsement of any specific regimen over another.
➢ The adjunctive use of mechanical compression devices is associated with a reduction in venous thromboembolism and no incremental bleeding risk, but data are not sufficient to recommend the use of mechanical compression devices as the sole means of venous thromboembolism prophylaxis at this time.
➢ Warfarin remains popular among North American surgeons for venous thromboembolism prophylaxis, largely because of its delayed onset of action and correspondingly low bleeding risk. However, it is not an ideal option because of the need for monitoring and unpredictable drug sensitivity in some patients that can result in occasional bleeding complications.
➢ Newer anticoagulants are very specific and more potent in their actions. They are highly effective in reducing deep venous thrombosis, but have not resulted in a meaningful decrease in clinical pulmonary embolism and are accompanied by a substantial increase in major and non-major clinically important bleeding.
➢ Recent observational data on aspirin prophylaxis suggest rates of clinical pulmonary embolism that are comparable with those observed with more potent anticoagulants, particularly when used in combination with regional anesthesia and pneumatic compression devices.
➢ The optimal regimen for the prevention of clinically important deep venous thrombosis and pulmonary embolism remains to be identified. Available evidence does not identify a clearly preferred best practice. A large pragmatic clinical effectiveness trial is warranted.
Elective total joint replacements of the hip and knee are collectively the most commonly performed surgical procedures in the United States, totaling nearly one million annually. Moreover, they are increasing in number with the aging of the baby boomers and constitute the single largest group of procedures covered by Medicare. Venous thromboembolism, specifically pulmonary embolism, is the most common cause of readmission and death after elective hip and knee replacement1-10. Despite advancements in perioperative care, fatal pulmonary embolism occurs in 0.1% to 0.5% of patients, accounting for >1000 deaths each year11-13. Appropriately, considerable effort has been directed at identifying the optimal method to prevent venous thromboembolism in these patients. However, the use of potent anticoagulants to mitigate the activation of thrombogenesis after orthopaedic operations must necessarily be tempered by the consideration of the bleeding risk following these procedures where hemostasis is imperfect in the setting of exposed osseous surfaces. Orthopaedic practitioners typically opt for one of three main regimens for perioperative chemoprophylaxis for venous thromboembolism: aspirin, low-intensity warfarin, or one of the newer anticoagulants such as fractionated heparins or direct factor-Xa or IIa inhibitors. Unfortunately, recommendations of professional associations have historically not been in agreement, and the ideal prophylaxis is yet to be determined. It must represent a risk-benefit balance in which the fear of fatal pulmonary embolism is weighed against the consideration for the morbidity of persistent wound drainage and hematoma14-16, secondary infection, and reoperation resulting from bleeding associated with perioperative anticoagulant use.
Development of Clinical Guidelines
It is not surprising that several federal agencies and public entities have weighed in over the years on the importance of venous thromboembolism as a public health issue. The National Institutes of Health (NIH) sponsored a consensus conference dedicated to venous thromboembolism in 198617 and concluded that, “for high-risk orthopedic patients undergoing elective hip surgery or knee reconstruction… low-dose warfarin, dextran, or adjusted-dose heparin” was “recommended… for at least 7 days” as prophylaxis. These regimens reduced the rate of clinical pulmonary embolism, but “the lowered death rate from PE [pulmonary embolism], while suggestive, [is] not statistically significant.” The panel also cautioned, “warfarin and dextran in commonly used doses can cause complications of operative bleeding and wound hematomas… [which] can be a significant problem in joint replacement patients”17. Over the subsequent three decades, venous thromboembolism prophylaxis of some type has become the standard of care after total hip arthroplasty and total knee arthroplasty in the United States. However, in 200318, the NIH Consensus Development Conference on Total Knee Replacement concluded that “the effectiveness of anticoagulation for the prevention of pulmonary emboli is unclear” and noted that prophylaxis recommendations were based “primarily on the reduction of deep venous thrombosis (DVT) detected by venography following TKR [total knee replacement]… and the available data indicate that DVT prophylaxis does not alter the occurrence of symptomatic DVTs or pulmonary embolism (PE), although no individual study was large enough to statistically assess effects on the occurrence of PE.” They concluded, “a randomized, placebo-controlled trial of prophylactic anticoagulation that assesses the outcomes of PE, bleeding, wound complications, and death seems warranted.”18
The American College of Chest Physicians (ACCP) took the lead with the development of clinical guidelines on this subject in 1992. They relied primarily on prospective, randomized clinical trials and had a relatively low level of concern for bleeding in deference to the principal intent of reducing the prevalence of deep venous thrombosis as a precursor and surrogate for pulmonary embolism19. After eight iterations of the ACCP Clinical Guidelines perennially endorsing potent anticoagulants for venous thromboembolism prophylaxis after total hip arthroplasty and total knee arthroplasty, including recommendation specifically against the use of aspirin, the American Academy of Orthopaedic Surgeons (AAOS) released its first set of clinical guidelines in 200920,21. Importantly, the AAOS considered observational data and emphasized clinical venous thromboembolism and death as primary end points with the intention of mitigating the risk of bleeding that may result in wound hematoma and secondary infection, both of which can require rehospitalization, reoperation, and removal of the prosthesis20. The AAOS favored less intensive anticoagulation regimens such as aspirin or low-intensity warfarin; both are associated with comparable rates of clinical pulmonary embolism and venous thromboembolism and lesser rates of bleeding than more potent anticoagulants recommended by the ACCP. Specifically, the AAOS endorsed the use of aspirin for patients at typical high risk of venous thromboembolism and introduced an algorithm of risk stratification for outliers. For patients perceived to be at an elevated risk of thrombosis, aspirin was dropped in favor of more potent anticoagulants. Conversely, for patients at an elevated risk of bleeding, potent anticoagulants were omitted in favor of aspirin, warfarin, or solitary use of mechanical compression. Notably, the AAOS recommendation for less potent anticoagulant prophylaxis in the face of an elevated risk of both venous thromboembolism and bleeding was identical to that for patients with only an elevated risk of bleeding and signaled that, from the surgeon’s perspective, an increased risk of adverse bleeding was at least as worrisome as an increased risk of fatal pulmonary embolism, which occurred substantially less often than did bleeding events. As such, clinical guidelines from the ACCP and AAOS remained at odds22. In 2008, the Surgeon General issued a Call to Action to Prevent Deep Vein Thrombosis and Pulmonary Embolism, based on the resulting frequent morbidity and mortality, which was estimated to account for at least 100,000 deaths each year23. In 2009, the Institute of Medicine released its national priorities for comparative effectiveness research and included “anticoagulant therapies (e.g., low-intensity warfarin, aspirin, injectable anticoagulants) for patients undergoing hip or knee arthroplasty surgery.”24
Guideline reconciliation finally occurred with the second edition of the AAOS guidelines in September 201125 and the ninth edition of the ACCP guidelines in February 201226, largely as a result of the agreement that clinically important pulmonary embolism or venous thromboembolism was the relevant end point and there were insufficient data to endorse any one specific prophylaxis regimen. Notably, both panels now comprise only individuals without relevant financial relationships with industry. The ACCP placed substantially greater value on the avoidance of bleeding based on acknowledgment of patient values, preferences, and aversion to surgical complications27. Likewise, in March 2012, the exhaustive AHRQ Comparative Effectiveness Review on Venous Thromboembolism Prophylaxis in Orthopedic Surgery suggested, “while the risk of developing deep vein thrombosis is highest followed by pulmonary embolism and major bleeding, there are inadequate data to say whether or not deep vein thrombosis causes pulmonary embolism or is an independent predictor of pulmonary embolism. The balance of benefits to harms is favorable for providing prophylaxis… and to extend the period of prophylaxis beyond the standard 7-10 days.”28 The report concluded that “interclass comparisons either could not be made due to lack of data, showed similarities between classes on outcomes, or had offsetting effects where benefits of one class on efficacy was tempered by an increased risk of bleeding.”28 Their first recommendation was that future comparative clinical trials should focus on final health outcomes rather than on a deep venous thrombosis surrogate. Likewise, it was suggested that future trials “should include more outcomes assessing harms such as bleeding leading to infection, bleeding leading to transfusion, readmission and reoperation to provide more information for the comparative balance of benefits to harms.”28 The Joint Commission29, Centers for Medicare & Medicaid Services30, and Surgical Care Improvement Project (SCIP)31 now mandate venous thromboembolism prophylaxis after total hip arthroplasty and total knee arthroplasty according to prevailing guidelines. In 2014, the SCIP added aspirin to its list of acceptable agents for prophylaxis.
By the consensus of the ACCP, AAOS, and AHRQ documents, all groups now decline to endorse any preferred prophylaxis and suggest only that something be done to reduce the risk of venous thromboembolism after joint replacement. Central to this impasse is an acknowledged absence of clinical pulmonary embolism and venous thromboembolism end-point data along with an increased appreciation of the morbidity associated with adverse bleeding and the patient’s perception of complications of anticoagulation prophylaxis27. These events occur at such a small rate that nearly 25,000 patients would be needed for a definitive clinical trial with an aggregate effectiveness end point of all-cause mortality and symptomatic pulmonary embolism and venous thromboembolism as the primary outcome measure. No existing studies satisfy this requirement of scale. As a result, specific clinical recommendations with regard to perioperative venous thromboembolism prophylaxis after hip and knee replacement remain elusive.
The adjunctive use of pneumatic compression devices to augment venous return and to increase fibrinolysis has become widespread in the orthopaedic community because these devices reduce deep venous thrombosis without an increase in bleeding12,17,18,20,32,33. Nevertheless, despite the recent availability of mobile compression devices34, the evidence does not support mechanical compression as the sole means of venous thromboembolism prophylaxis, particularly relative to the prevention of proximal deep venous thrombosis, in this high-risk group at this time35-37. One observational registry of 3060 nonrandomized patients using mobile compression demonstrated noninferiority relative to published chemoprophylaxis results but was funded by the device manufacturer, did not control for the concurrent use of aspirin, and was subject to selection bias among surgeon investigators38. The variable compliance, unproven effectiveness, and impracticality of walking devices remain impediments to adoption of pneumatic compression as the sole method of prophylaxis. Less cumbersome devices and wear-time monitors that prompt use may improve compliance and effectiveness. These present limitations support the continued search for the ideal chemoprophylaxis that balances effective pulmonary embolism prevention with acceptable bleeding risk.
Current chemoprophylaxis can be grouped into three main categories: warfarin, aspirin, and newer selective heparinoids or direct factor-Xa or IIa inhibitors.
Low-intensity warfarin (International Normalized Ratio [INR] target, 2.0), despite its variable dosing and need for monitoring39, is a time-honored option for orthopaedic thromboprophylaxis, likely because it has an acceptably low bleeding risk (1% to 2%) coupled with good efficacy in preventing clinically important pulmonary embolism40. In surveys conducted in 2000 and again in 2008, it was the venous thromboembolism prophylaxis of choice for nearly 50% of all orthopaedic surgeons performing total joint arthroplasty prior to the release of direct factor-Xa and thrombin inhibitors in the United States41,42. A survey of 465 members (representing a 55% response rate) of the American Association of Hip and Knee Surgeons (AAHKS) noted that up to 47% used warfarin, 28% used low-molecular-weight heparin, and nearly 20% preferred aspirin42. Also in 2008, a mail survey of 3025 general orthopaedic surgeons yielded 634 respondents (21%) and revealed an association between performing more joint replacements and a progressively smaller likelihood of using fractionated heparin for venous thromboembolism prophylaxis43. Among surgeons performing fewer than sixty-four arthroplasties annually, 75% used low-molecular-weight heparin, compared with only 57% among surgeons performing more than 140 arthroplasties annually43. Warfarin remains popular in the United States, despite the inconvenience and expense of monitoring, because many believe that it represents the best available compromise of efficacy in preventing clinical venous thromboembolism and safety in minimizing bleeding.
Warfarin is an orally administered vitamin-K antagonist that blocks the synthesis of related procoagulant factors (II, VII, IX, and X) and physiologic anticoagulants (proteins C and S) in the liver. Its effect lags administration by forty-eight hours and is readily reversible by administration of fresh frozen plasma or parenteral vitamin K. Critics have been most concerned with its erratic effect, need for monitoring, and associated bleeding. Early studies showed bleeding rates of 10% with a prothrombin time index of 1.8 to 2.0, but the equivalent efficacy is now known to occur with a prothrombin time index of 1.3 to 1.5 (INR, 2.0) and reduced bleeding to <2%39. Historically, low-intensity warfarin prophylaxis (INR, 2.0) has residual venographic deep venous thrombosis rates of 9% to 26% and proximal clot rates of 2% to 5% after total hip arthroplasty40,44,45 compared with overall deep venous thrombosis rates of 35% to 55% and proximal clot rates of 2% to 14% after total knee arthroplasty40,44. The prevention of deep venous thrombosis after total knee arthroplasty has been far more recalcitrant to chemoprophylaxis than after total hip arthroplasty, especially when using warfarin. After total hip arthroplasty, warfarin preferentially reduces proximal compared with distal deep venous thrombosis35 and, when combined with forty-eight hours of intraoperative epidural anesthesia and postoperative epidural analgesia, the residual deep venous thrombosis rate dropped below 10%46. Most importantly, extended low-intensity warfarin prophylaxis continued for six weeks after total hip arthroplasty and total knee arthroplasty in 3293 patients was associated with readmission for clinical venous thromboembolism in 0.3% (one of 360 patients) after total hip arthroplasty7 and 0.2% (one of 484 patients) after total knee arthroplasty8 with a major bleeding rate of <0.1% (three of 3293 patients). In the combined population of patients receiving warfarin after undergoing total hip arthroplasty (n = 360) or total knee arthroplasty (n = 484) from those studies7,8, six weeks of warfarin after hospital discharge eliminated the risk of pulmonary embolism (none of the 844 patients in the warfarin group compared with seventeen of the 2449 patients in the group that did not receive extended warfarin prophylaxis; p = 0.01) and significantly reduced venous thromboembolism-related readmissions (0.2% [two of 844 patients, one in each of the total hip arthroplasty and total knee arthroplasty cohorts] in the warfarin group compared with 1.6% [thirty-eight of 2449 patients] in the group that did not receive warfarin; p = 0.0015)7,8,47.
The optimal use of warfarin remains challenging and labor-intensive; it begins the evening before surgery because of the forty-eight-hour latency of its anticoagulant effect. Indeed, this lag in activity is likely responsible for its popular safety margin with low perioperative bleeding rates. The combination of preoperative warfarin and epidural anesthesia is safe and effective and further reduces residual deep venous thrombosis by 50% compared with general anesthesia12,13,46. Moreover, extended warfarin after discharge prevents venous thromboembolism-related morbidity and mortality even in patients who developed deep venous thrombosis despite primary warfarin prophylaxis7,8,48-50. Nevertheless, an inconsistent dose response leading to untoward bleeding, the need for vigilant INR monitoring and dose adjustment, and the associated costs all fuel the search for a more predictable oral anticoagulant that retains a favorable safety profile without monitoring.
Aspirin, acetylsalicylic acid, is traditionally considered an arterial circulation drug and has been controversially used in conjunction with mechanical compression devices for venous thromboembolism prophylaxis by a consistent minority (10% to 20%) of surgeons. Conventional wisdom holds that aspirin reduces arterial thrombosis through inhibition of platelet cyclooxygenase-1, which decreases synthesis of thromboxane A2 (platelet-activating eicosanoid) and related platelet activation, and aspirin has demonstrated substantial arterial thrombotic event reduction as well as survival benefit relative to stroke, myocardial infarction, and related deaths in high-risk patients51,52. Conversely, clotting experts often regard aspirin as an inferior and inconsistent agent in mitigating venous thrombosis53. A 1986 NIH consensus conference concluded that aspirin had no role in venous thromboembolism prophylaxis after total hip arthroplasty and total knee arthroplasty17, and this sentiment was perpetuated in the ACCP guidelines until only recently19.
Since 2006, three large observational studies54-56 and a national joint registry study57 have rekindled the notion that aspirin can prevent clinical pulmonary embolism after total joint arthroplasty, especially in conjunction with regional anesthesia. More than 2000 patients undergoing total hip arthroplasty were managed with hypotensive epidural anesthesia, intravenous heparin during femoral preparation, and in-hospital pneumatic compression, followed by six weeks of aspirin (325 mg twice a day) or low-intensity warfarin (18% of patients)54. The overall venous thromboembolism-related readmission rate for total hip arthroplasty was 3.2%, with a clinical pulmonary embolism rate of 0.6% and no fatal pulmonary embolism. In a second study, more than 3400 total knee arthroplasties were performed with use of regional anesthesia followed by aspirin (325 mg twice a day) or warfarin (2% of patients) for six weeks55. The overall venous thromboembolism-related readmission rate for total knee arthroplasty was 0.5%, with a clinical pulmonary embolism rate of 0.36% inclusive of 0.1% fatal pulmonary embolism. In the third study, a combination of consecutive primary total hip arthroplasties (n = 2203) and total knee arthroplasties (n = 2050) were performed with use of spinal anesthesia, followed by aspirin (150 mg daily) or anticoagulants (2% of patients) for six weeks56. The combined overall venous thromboembolism-related readmission rate for total hip arthroplasty and total knee arthroplasty was 1.1% (forty-five patients), with a clinical pulmonary embolism rate of 0.73% (thirty-one patients) inclusive of 0.07% fatal pulmonary embolism (three patients). Finally, 2009 National Joint Registry data suggested that 20% of surgeons in the United Kingdom used aspirin as primary venous thromboembolism prophylaxis after total hip arthroplasty and total knee arthroplasty57. A cohort subgroup analysis of 22,942 patients receiving aspirin matched with the same number receiving low-molecular-weight heparin revealed no difference in ninety-day outcomes related to clinical pulmonary embolism (0.7%), deep venous thrombosis (0.95%), stroke or gastrointestinal bleeding (0.75%), or reoperation (0.35%). However, patients who received low-molecular-weight heparin had a survival advantage, with all-cause mortality of 0.49% compared with 0.65% for those who received aspirin (relative risk = 0.75; p = 0.02).
Finally, a reassessment of the Pulmonary Embolism Prevention (PEP) trial, a study comparing aspirin with placebo use in 13,356 patients with hip fracture and 4088 patients undergoing elective total hip arthroplasty or total knee arthroplasty with a thirty-five-day treatment and clinical pulmonary embolism or mortality outcome window, effectively prompted the inclusion of aspirin in the ACCP clinical guidelines58. Its original interpretation in 2000 noted a 34% reduction in overall venous thromboembolism with aspirin only in the hip fracture group, but a more recent pragmatic interpretation as a clinical effectiveness trial endorsed an overall 19% reduction (1.15% compared with 1.4%) in aggregate venous thromboembolism end points in the combined group of 17,444 patients undergoing either hip fracture repair or elective arthroplasty.
The U.S. Food and Drug Administration (FDA) approval of oral direct factor-IIa and Xa inhibitors that have greater perioperative bleeding risk has stimulated an increase in use of aspirin (given its lower bleeding rates and recent acceptance by the ACCP and SCIP) for venous thromboembolism prophylaxis by arthroplasty surgeons. Moreover, in 2012, two important randomized clinical trials were published concerning aspirin use to prevent recurrent venous thromboembolism. In both the Aspirin for the Prevention of Recurrent Venous Thromboembolism [Warfarin and Aspirin (WARFASA)]51,59 study and the Aspirin to Prevent Recurrent Venous Thromboembolism (ASPIRE)52,60 trial, patients with unprovoked venous thromboembolism within two years of enrollment were randomized to either aspirin (100 mg daily) or placebo after completion of an initial course of standard warfarin anticoagulation. In 402 patients in the WARFASA study, aspirin reduced the risk of unprovoked recurrent venous thromboembolism from 11.0% to 5.9% (hazard ratio, 0.55; p = 0.02) over a median two-year treatment59. In 822 patients in the ASPIRE trial, aspirin reduced the risk of unprovoked recurrent venous thromboembolism from 6.5% to 4.8% (hazard ratio, 0.74; p = 0.09) over a median thirty-seven-month follow-up60. Patients in the ASPIRE trial receiving aspirin enjoyed a reduction in composite major vascular events (overall venous thromboembolism, myocardial infarction, stroke, or cardiovascular death) from 8.0% to 5.2% per year (hazard ratio, 0.66; p = 0.01). Patients receiving aspirin experienced an overall net clinical benefit (aggregate of recurrent venous thromboembolism, myocardial infarction, stroke, major bleeding, and all-cause mortality) with an annual event rate reduction from 9.0% to 6.0% (hazard ratio, 0.67; p = 0.01)60. Considering these two trials together, aspirin was associated with a 32% reduction in recurrent venous thromboembolism (hazard ratio, 0.68, p = 0.007) and a 34% reduction in major adverse vascular events (hazard ratio, 0.66; p = 0.002), without any accompanying increase in bleeding51,52. Given its efficacy without a compromise in safety, although acknowledging that these trials involved spontaneous recurrence of deep venous thrombosis rather than that provoked by a potent surgical stimulus, these two studies have legitimately stimulated a reassessment of the role of aspirin for venous thromboembolism prophylaxis and the specific prevention of embolization of existing venous thrombi. However, just as less intensive anticoagulation was gaining in popularity, a recent prospective observational trial comparing aspirin with warfarin using AAOS guidelines and risk stratification for venous thromboembolism prophylaxis after total hip arthroplasty and total knee arthroplasty was discontinued prematurely because of failure of aspirin prophylaxis61. Acute symptomatic venous thromboembolism occurred in 1.2% (five of 415 patients) on warfarin or another ACCP-endorsed drug compared with 7.9% (twelve of 152 patients) in the standard-risk group receiving aspirin (relative risk, 7.0; p = 0.001). Comparing all patients specifically receiving warfarin in either group (537 patients) with the patients receiving aspirin (152 patients), clinical pulmonary embolism occurred in 0.6% (three patients) in the warfarin group and 4.6% (seven patients) in the aspirin group (p = 0.01). All clinical venous thrombolic events occurred in 1.1% (six patients) in the warfarin group and 7.9% (twelve patients) in the aspirin group (p = 0.001). The authors suggested that either aspirin was inferior to warfarin prophylaxis or the AAOS risk stratification tool did not adequately discriminate between patients at standard risk and those at elevated risk61. The inadequacy of the risk stratification protocol underscores the importance of randomization in such prospective trials where avoidance of selection bias is critical to ensure clinical equipoise.
In the nearly thirty years since the NIH conference on venous thromboembolic disease, many new selective anticoagulant drugs have been introduced, but the evidence for the prevention of fatal pulmonary embolism after total hip arthroplasty and total knee arthroplasty has changed very little47,62. Randomized clinical trials have demonstrated a dramatic reduction in venographic thrombi without a parallel reduction in fatal pulmonary embolism. Fractionated heparins63-65, synthetic pentasaccharide66, and direct factor-Xa10,67-69 and thrombin (IIa) inhibitors70,71 all offer substantially improved efficacy in reduction of venographic thrombosis but are uniformly associated with increased perioperative bleeding risk72-74. Despite residual venographic clot rates up to five times greater than with these newer agents, low-intensity warfarin (INR, 2.0) and aspirin offer comparable clinical pulmonary embolism rates and a twofold to threefold reduction in major bleeding complications6-8,10.
Low-molecular-weight (fractionated) heparins and pentasaccharide are progressively smaller sugar molecules with increasing specificity in binding antithrombin III (AT-III) as the critical intermediary prior to deactivating direct factor-X or IIa (thrombin). Indeed, synthetically derived pentasaccharide is ingeniously made to exactly correspond to the five-sugar AT-III binding site and has enhanced specificity to bind factor Xa. Because of the more proximal location of factor Xa in the coagulation cascade, this affinity for factor Xa greatly augments the potency of low-molecular-weight heparins and pentasaccharide. Accordingly, low-molecular-weight heparin reduced overall venographic deep venous thrombosis from 44% to 31% after total hip arthroplasty compared with unfractionated heparins63; deep venous thrombosis reduction has also been substantial after total knee arthroplasty, but less dramatic. Comparison with warfarin is less compelling; one study involving two 1500-patient cohorts noted identical all-cause mortality (0.8%) and comparable overall venous thromboembolism (3.6% compared with 3.8%) three months postoperatively, but a nearly threefold increase in clinical bleeding events (1.3% compared with 0.5%) with enoxaparin compared with warfarin64. Similarly, in 1472 patients undergoing total hip arthroplasty, preoperative dalteparin compared with warfarin reduced symptomatic deep venous thrombosis (1.5% compared with 4.4%; p = 0.02) but increased major bleeding events (8.9% compared with 4.5%; p = 0.01) and surgical site bleeding events (8.3% compared with 3.9%; p = 0.03)65. Fondaparinux (pentasaccharide) further reduced venographic deep venous thrombosis by 50% compared with enoxaparin after total hip arthroplasty, total knee arthroplasty, and hip fracture. It was the first agent to reduce deep venous thrombosis to <20% after total knee arthroplasty and the most extensively studied agent after hip fracture, but its bleeding rate of 3% to 6% exceeded even that seen with enoxaparin66. One meta-analysis demonstrated a 50% increase in bleeding after total hip arthroplasty with low-molecular-weight heparin compared with warfarin prophylaxis73.
New direct factor-X and IIa inhibitors require no monitoring and are administered orally; however, no antidotes are currently available. Rivaroxaban, a direct factor-Xa inhibitor, has been most extensively studied in two total hip arthroplasty trials (7050 patients) and two total knee arthroplasty trials (5679 patients)10,67-69. Compared with enoxaparin, a pooled analysis of the four studies demonstrated a 58% reduction in all-cause mortality and symptomatic venous thromboembolism (0.6% compared with 1.3%; p < 0.001), but aggregate bleeding data revealed an increase in major bleeding events and clinically important non-major bleeding events (3.2% compared with 2.6%; p = 0.039)10,67-69. Rivaroxaban has enjoyed rapid growth in popularity in the venous thromboembolism prophylaxis market, but an increase in bleeding events has dampened enthusiasm for its widespread adoption by the orthopaedic community.
Likewise, one observational study demonstrated that the all-cause mortality after total hip and knee replacement in patients who had been given potent anticoagulants was more than twice that in patients receiving aspirin, pneumatic compression, and regional anesthesia75. Patients receiving fractionated heparin developed nonfatal pulmonary embolism 60% to 70% more often than those receiving only aspirin and mechanical devices, indicating that pulmonary embolism occurs despite prophylaxis with potent anticoagulants75 and should not be considered a “never event.”23,29,30 Another report specifically noted the failure of low-molecular-weight heparin compared with warfarin prophylaxis; symptomatic deep venous thrombosis (3.8%), nonfatal pulmonary embolism (1.3%), persistent wound drainage resulting in readmission (4.7%), and reoperation (3.4%) all occurred at rates exceeding prior experience with low-intensity warfarin72. Moreover, two recent reports from the United Kingdom during a period when hospital trusts directed discontinuation of low-molecular-weight heparins in favor of adoption of rivaroxaban have focused attention on the issue of perioperative bleeding with rivaroxaban after total hip and knee replacement76,77. In one series, there was no difference in all-cause mortality, pulmonary embolism, and major bleeding events. Symptomatic deep vein thrombosis was reduced from 0.91% in 10,361 patients receiving low-molecular-weight heparin to 0.36% in 2762 patients receiving rivaroxaban, but wound complications increased from 2.81% in the low-molecular-weight heparin group to 3.85% in the rivaroxaban group (p = 0.005)76. Finally, one retrospective review of 1048 consecutive patients undergoing total hip arthroplasties or total knee arthroplasties showed a twofold increase (p = 0.046) in thirty-day reoperation for wound complications with rivaroxaban at 3.94% (twenty-two of 559 patients) compared with tinzaparin prophylaxis at 1.8% (nine of 489 patients)77. Not surprisingly, the orthopaedic community has been slow to adopt these newer agents in favor of a more balanced strategy that offers less bleeding risk with comparable protection against clinical venous thromboembolism disease.
Current Best Practice, Public Health, and Future Directions
Arguably fueled by intense public scrutiny of physician conflict of interest on review panels, increasing respect for patient-perceived adverse bleeding events27, and reassessment of the role of aspirin in venous thromboembolism prevention, current AAOS and ACCP clinical guidelines and the AHRQ Effectiveness Review are in nearly complete agreement25,26,28. Each concluded that the paucity of data on clinical venous thromboembolism outcomes did not justify the endorsement of any preferred regimen, and all include aspirin among an expanded list of acceptable prophylaxis options. Although this generic convergence of opinion provides welcome resolution to the medicolegal conundrum for practitioners, it provides little guidance as to clinical best practice (Table I).
At a 2011 multidisciplinary international thrombosis workshop, expert consensus was that true clinical equipoise existed on this subject, resulting from too little clinical data and the limitations of available observational administrative data sets78. The present state creates an impasse that ethically supports a unique opportunity for conduct of a large randomized clinical trial that will change current practice and inform future treatment decisions. The ideal prophylaxis remains elusive; it must represent a balance among the risk of death from pulmonary embolism and major hemorrhage, the morbidity of bleeding associated with anticoagulation, and the preferences of individual patients. If aspirin were found to be comparable with rivaroxaban in all-cause mortality and preventing clinically meaningful venous thromboembolism, the drug cost savings alone for thirty-five days of prophylaxis according to prevailing guidelines would be in excess of $400 million for one million procedures performed annually in the United States79. Considering the additional societal cost attributable to increased bleeding complications and related reoperation with the newer anticoagulants, there is compelling justification to conduct a definitive clinical trial to ascertain the optimal regimen for venous thromboembolism prophylaxis after total hip arthroplasty and total knee arthroplasty. Indeed, we cannot afford to not conduct such a clinical trial that is so essential to the clinical well-being of our patients and economically critical to our health-care system.
Source of Funding: There was no source of external funding for this study.
Investigation performed at the Department of Orthopaedics, Medical University of South Carolina, Charleston, South Carolina
Disclosure: The author did not receive payments or services, either directly or indirectly (i.e., via his institution), from a third party in support of any aspect of this work. Neither the author nor his institution has had any 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. The author has a patent or patents, planned, pending, or issued, that is broadly relevant to the work. The author has not had any other relationships, or 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 © 2015 by The Journal of Bone and Joint Surgery, Incorporated