The Safety and Efficacy of Extended Thromboprophylaxis With Fondaparinux After Major Orthopedic Surgery of the Lower Limb With or Without a Neuraxial or Deep Peripheral Nerve Catheter: The EXPERT Study

Abstract

BACKGROUND: 

The benefit-risk ratio of extended fondaparinux therapy has not been assessed in patients undergoing major lower limb joint arthroplasty. Few data on the concomitant use of fondaparinux and continuous neuraxial or deep peripheral nerve blockade are available. We performed a prospective intervention study in patients undergoing major orthopedic surgery primarily designed to assess the efficacy of fondaparinux when drug administration was withheld for 48 h to permit removal of a neuraxial or deep peripheral nerve catheter. The safety and efficacy of extended fondaparinux therapy for the prevention of venous thromboembolism were also evaluated.

METHODS: 

Patients received a daily subcutaneous injection of 2.5 mg fondaparinux for 3 to 5 wk postoperatively. In patients with a neuraxial or deep peripheral nerve catheter, the catheter was removed 36 h after the last fondaparinux dose. The next fondaparinux dose was administered 12 h after catheter removal. The primary end points were symptomatic venous thromboembolism and major bleeding up to 4–6 wk after surgery.

RESULTS: 

We recruited 5704 patients. A neuraxial or deep peripheral nerve catheter was inserted in 1553 (27%) patients and 78 (1.4%) patients, respectively. The rate of venous thromboembolism was 1.0% (54 of 5387). There was no difference between patients without (1.1%) or with (0.8%) a catheter (the upper limit of the 95% confidence interval of the odds ratio, 1.49, being below the predetermined noninferiority margin of 1.75). The incidence of major bleeding was 0.8% (42 of 5382). No neuraxial or perineural hematoma was reported.

CONCLUSIONS: 

Once-daily subcutaneous injection of 2.5 mg fondaparinux given for 3 to 5 wk was effective and safe for prevention of venous thromboembolism after major orthopedic surgery. Temporary discontinuation of fondaparinux for 48 h permitted safe removal of a neuraxial or deep peripheral nerve catheter without decreasing thromboprophylatic efficacy.

In Brief

IMPLICATIONS: This observational study substantiated the clinical value of once-daily 2.5 mg fondaparinux given for 4 ± 1 wk after major orthopedic lower limb surgery. The clinical value of fondaparinux was also observed in patients in whom specific recommendations for neuraxial/deep peripheral indwelling catheter removal were applied.

Fondaparinux is a selective inhibitor of factor Xa. In four phase III clinical studies in 7344 patients undergoing major orthopedic surgery, a once-daily subcutaneous injection of 2.5 mg fondaparinux administered postoperatively for 5 to 9 days reduced the incidence of venographic venous thromboembolism by more than 50% compared with twice-daily 30 mg or once-daily 40 mg enoxaparin without increasing the incidence of clinically significant bleeding (^1–5). Four-week fondaparinux administration after hip fracture surgery reduced the incidence of venous thromboembolism by 90% compared with a 1-wk fondaparinux treatment followed by 3-wk placebo (⁶). Therefore, in 2004, the American College of Chest Physicians recommended, among other anticoagulant drugs, once-daily subcutaneous injection of 2.5 mg fondaparinux for at least 10 days in patients undergoing elective knee arthroplasty, and for 28 to 35 days in patients undergoing elective hip arthroplasty or hip fracture surgery (⁷). Nevertheless, additional data on the clinical value of extended fondaparinux prophylaxis are still required, especially in patients undergoing joint arthroplasty surgery.

Neuraxial anesthesia/analgesia or deep peripheral (i.e., lumbar plexus or parasacral sciatic) nerve blockade is frequently used to provide postoperative analgesia after major orthopedic surgery of the lower limb. These procedures are associated with a risk of epidural, spinal, or perineural hematoma, which can induce long-term or permanent nerve paralysis, especially in patients concomitantly receiving anticoagulant drugs (⁸). No neuraxial hematoma has been reported in the previous major orthopedic and abdominal surgery trials with once-daily injections of 2.5 mg fondaparinux (^1–4,6,9). However, the safety of the concomitant use of fondaparinux and continuous neuraxial or deep peripheral nerve blockade needs to be investigated in routine practice.

The primary objective of the study in patients undergoing major orthopedic lower limb surgery was to compare the safety and efficacy of fondaparinux in patients with a neuraxial or deep peripheral catheter (i.e., patients in whom the drug administration was discontinued for 48 h to permit safe catheter removal) with the safety and efficacy in patients without a catheter. The study also evaluated the safety and efficacy of extended fondaparinux therapy for the prevention of symptomatic venous thromboembolism.

METHODS

The EXPERT trial was an international (19 countries), multicenter (243 centers), nonrandomized, open-label, prospective, intervention study (Appendix available at www.anesthesia-analgesia.org). The study was conducted according to the ethical principles as stated in the Declaration of Helsinki and local regulations. The protocol was approved by independent ethics committees and written informed consent was obtained from all patients.

Study Population

Patients older than 18-yr-of-age undergoing major orthopedic surgery of the lower limb and requiring at least 3 wk of thromboprophylaxis were eligible for the study, regardless of the technique for anesthesia and postoperative analgesia. Patients were excluded from the study if unusual difficulties were encountered in performing epidural or spinal anesthesia (i.e., patients in whom more than three attempts were required and/or the puncture was complicated by bleeding); if a neuraxial or deep peripheral nerve (lumbar plexus and/or parasacral sciatic nerve) catheter was placed or planed and the patient required antiplatelet drugs (aspirin, ticlopidine, or clopidogrel) or received ticlopidine or clopidogrel in the last 7 days, or aspirin in the last 3 days, or if the plan was to withdraw a neuraxial or deep peripheral catheter the day after surgery. Patients fulfilling at least one of the following criteria were also excluded from the study: acute bacterial endocarditis; active bleeding; known congenital or acquired bleeding disorder; current ulceration or angiodysplastic gastrointestinal disease; hemorrhagic stroke; brain, spinal or ocular surgery within the previous 6 mo; other elective surgery scheduled within the next 6 wk; pregnancy; lactation; serum creatinine concentration above 194 μmol/L (2.2 mg/dL); creatinine clearance below 30 mL/min; platelet count below 100 × 10⁹/L; need for anticoagulant drugs for any chronic condition; administration of either thrombolytic drugs or platelet glycoprotein IIb/IIIa receptor antagonists within 3 days before surgery, or any type of anticoagulant or antiplatelet drugs closer to surgery than recommended in local guidelines; or hypersensitivity to fondaparinux.

Patients were segregated into “catheter” or “no catheter” group depending on whether a neuraxial or deep peripheral nerve catheter was placed. The anesthetic plan was not randomized, and was entirely left to the discretion of the attending anesthesiologist.

Study Design

Patients received once-daily subcutaneous injections of 2.5 mg fondaparinux (Arixtra®, GlaxoSmithKline, Research Triangle Park, NC) for 3 to 5 wk after surgery according to the local prophylactic practices. The first dose was administered between 6 and 12 h after surgical closure (Fig. 1). All patients were followed for 4 to 6 wk after surgery.

Figure 1.: Design of the EXPERT study.

Specific recommendations were established for catheter removal in patients with a neuraxial or deep peripheral nerve (lumbar plexus, or parasacral sciatic nerve) catheter (Fig. 2). Fondaparinux was discontinued for 48 h to permit safe catheter removal. Because the risk of hematoma is highest when the catheter is removed and the anticoagulant activity is high (⁹), it was recommended not to withdraw the catheter until 36 h after the last fondaparinux dose. This is well after the T_max of 2 h after subcutaneous injection (¹⁰), and approximately twice the 17-h half-life of fondaparinux (¹¹). This is why patients could not be enrolled if catheter withdrawal was planned for the day after surgery. In practice, fondaparinux administration was skipped the day before catheter removal, the catheter was withdrawn 36 h after the last fondaparinux administration, and the next fondaparinux dose was administered 12 h after catheter removal. If the catheter had to be removed for an urgent reason, it was withdrawn as late as possible after the last dose, and the next fondaparinux dose was administered 24 h after catheter withdrawal. In the case of a spontaneous catheter withdrawal, the next fondaparinux dose was given the following day as planned. Patients were followed with a careful neurological examination for 24 h after catheter removal. No specific recommendations were given concerning the management of superficial (e.g., femoral, popliteal, or sciatic nerve) catheters.

Figure 2.: Recommendation for epidural/deep peripheral catheter withdrawal.

Medications

Fondaparinux was administered using a prefilled, single-dose syringe containing 2.5 mg of fondaparinux sodium in 0.5 mL of water. Throughout the treatment period, unfractionated heparin, low-molecular-weight heparin, vitamin K antagonist, dextran, platelet glycoprotein IIb/IIIa receptor antagonist, or any thrombolytic therapy was not allowed. The use of nonsteroidal antiinflammatory drugs (NSAIDs) was discouraged. If their use was necessary, selective cyclooxygenase-2 (COX-2) inhibitors or short half-life NSAIDs were preferred. Additionally, if NSAIDs or COX-2 inhibitors were used, the highest dosages were avoided, they were started postoperatively, and their use contraindicated the concomitant administration of any antiplatelet drugs. The administration of antiplatelet drugs was also discouraged. If an antiplatelet drug was required, it contraindicated the concomitant administration of any NSAID or selective COX-2 inhibitor. Furthermore, antiplatelet and NSAID use was excluded in patients with a neuraxial or deep peripheral nerve catheter. In this case, the antiplatelet treatment had to be stopped at least (depending on local guidelines) 3 (aspirin) to 7 (thienopyridines) days before inclusion in the trial. The use of elastic stockings and intermittent pneumatic compression was allowed.

Study End Points

The primary safety outcome was major bleeding up to 4–6 wk after surgery. Major bleeding was defined as fatal bleeding, bleeding in a critical organ (i.e., retroperitoneal, intracranial, spinal, intraocular, or pericardial), bleeding at the surgical site leading to reoperation, or bleeding at a nonsurgical site requiring the transfusion of more than two units of blood. Secondary safety outcomes were death, and any other adverse events.

The primary efficacy end point was symptomatic venous thromboembolism up to 4–6 wk after surgery in both groups. Venous thromboembolism was defined as symptomatic deep vein thrombosis (confirmed by ultrasonography or venography) or pulmonary embolism (confirmed by ventilation/perfusion scanning, pulmonary angiography, or spiral computed tomography, or documented on appropriate medical records if fatal).

Efficacy outcomes, bleeding and death were validated by an external independent adjudication committee, the members of which were not aware of whether or not the patient had received a neuraxial or deep nerve catheter.

Statistical Analysis

The primary objective of the study was to assess the impact on the rate of symptomatic venous thromboembolism if the fondaparinux was discontinued for 48 h to permit safe removal of a neuraxial or deep peripheral nerve catheter. Preserving at least 65% of the effect of fondaparinux versus placebo (odds ratio [OR]: 0.11 [95% confidence interval: 0; 0.32]) (⁶) and considering the upper limit of the 95% confidence interval, we calculated that the rate of symptomatic venous thromboembolism in patients with a catheter would be noninferior to that in patients without a catheter if the upper limit of the confidence interval of the OR between patients with and without catheter on the primary efficacy end point was below 1.75. Assuming a venous thromboembolism incidence 5 wk after surgery of approximately 1.5% (^5,6) and the insertion of a neuraxial or deep peripheral nerve catheter in up to one-third of patients, it was calculated that 6500 patients were needed to meet the primary objective of the study with a one-sided α risk of 5%, a power of 80%, and a relative noninferiority margin of 75%. Because the recruitment process was slower than expected, and neuraxial catheters for postoperative analgesia were replaced at the majority of centers by superficial peripheral nerve catheters, the Steering Committee, unaware of the results, discontinued the study prematurely after the enrollment of 5713 patients.

All analyses were performed using SAS software (Version 8.2, SAS Institute, Cary, NC). All patients having received at least one study drug were included in the intention-to-treat (ITT) population. All ITT patients without protocol violations were included in the per-protocol (PP) population. Protocol violations were defined before study initiation as patients who had not given their informed consent, who received fondaparinux on the day of catheter removal, or patients without a catheter who did not receive one or more scheduled fondaparinux doses in the first 5 days after surgery. The effect of withholding the drug administration for 48 h on the efficacy of fondaparinux was determined by calculating the OR for venous thromboembolism between patients with and without a catheter. Noninferiority was concluded if the upper limit of the 95% confidence interval of the OR did not reach the predefined noninferiority margin of 1.75.

The robustness of the primary efficacy analysis was assessed by performing sensitivity analyses according to the following three scenarios for handling patients with a missing evaluation of symptomatic venous thromboembolism. The “best case” scenario assumed that none of these patients experienced venous thromboembolism. The “realistic case” scenario applied the highest observed rates of venous thromboembolism to the missing patients in either group of patients, with and without catheter. In the “worst case” scenario, all of the missing patients were assumed to have experienced venous thromboembolism.

Furthermore, because this trial was not randomized, a difference in venous thromboembolism risk between patients with or without a catheter was possible. Therefore, we performed a multivariate logistic analysis including a propensity score adjustment which considered a possible difference in covariates between patients with and without a catheter (¹¹).

RESULTS

Patients

We recruited 5713 patients between July 2003 and October 2004. Nine patients did not receive the study drug leaving 5704 (99.8%) patients available (Fig. 3). Overall, 94% and 90% of the recruited patients were available for the ITT and PP analyses, respectively. In the ITT efficacy population, the rate of withdrawal was 5.5% (222 of 4074) in patients without a catheter and 5.8% (95 of 1630) in patients with a catheter. The two main protocol violations were the administration of fondaparinux on the day of catheter removal in 112 patients, and the failure to administer of one or more scheduled doses of fondaparinux in the 5 days after surgery in 160 patients without a catheter.

Figure 3.: Trial profile. Of note, no bleeding was reported in the 112 patients in whom fondaparinux was administered on the day of catheter removal. Three venous thromboembolic episodes were reported in the 160 patients without a catheter in whom one or more scheduled doses of fondaparinux were not administered in the 5 days after surgery.

The mean age of the population was 66 yr; 24% of patients being at least 75-yr-old (Table 1). There were twice as many women as men. Obesity (body mass index ≥30 kg/m²) was observed in 30% of the patients, and moderate or severe renal impairment (creatinine clearance <50 mL/min) in 10%. Approximately 5% of patients had a history of venous thromboembolism. Overall, the baseline demographic and medical characteristics of the patients were similar, regardless of whether or not they had a catheter.

Table 1: Population Characteristics

Surgery and Treatment

The main results are presented in Table 2. Fifty-two percent of the patients underwent total hip replacement, 40% knee replacement, and 6% hip fracture surgery (all proximal femoral fractures). Many operations (62%) were performed under regional anesthesia only. A neuraxial catheter was placed in 27% of patients, a deep peripheral nerve block catheter in 1.4%, and a superficial (i.e., femoral or popliteal sciatic) nerve block catheter in 7% (405 of 5704). The neuraxial or deep peripheral nerve block catheter was removed between 1 and 2 days after surgery in 43% (706 of 1626) of patients, and between 3 and 6 days after surgery in 57% (920 of 1626). Fondaparinux was initiated between 6 and 12 h after surgery in 95% (5101 of 5376) of patients, <6 h after surgery in 2% (108 of 5376) of patients, and more than 12 h after surgery in 3% (167 of 5376) of patients. Fondaparinux was given for a median of 35 (range: 1–105) days. More than 92% of patients received fondaparinux for more than 3 wk, and 84% at least 4 wk. Elastic stockings were given to 66% of patients. NSAIDs were given to 57% of patients. Patients were discharged after a mean period of 10.8 days. The above data did not differ substantially between patients with and without a catheter.

Table 2: Surgery and Treatment Characteristics

Clinical Outcomes

In the ITT population, the rate of adjudicated symptomatic venous thromboembolism at 4 to 6 wk after surgery was 0.8% in patients with a catheter and 1.1% in patients without a catheter (OR: 0.79 [95% confidence interval: 0.42 to 1.49]). In the PP population, the rate of adjudicated symptomatic venous thromboembolism was 0.8% (12 of 1428) in patients with a catheter and 1.0% (38 of 3708) in patients without a catheter (OR: 0.82 [95% confidence interval: 0.43 to 1.57]). In both analyses, the upper limit of the 95% confidence interval was below the predetermined noninferiority margin of 1.75, demonstrating that the efficacy of fondaparinux in patients with a catheter in whom the drug was not administered for 48 h was not inferior to that observed in patients without a catheter.

On logistic regression analysis, venous thromboembolism rates were similar in patients with or without a catheter regardless of various covariates, including gender, age, body mass index, creatinine clearance, a history of venous thromboembolism, type of surgery, type of anesthesia, and treatment duration (data not shown). Furthermore, noninferiority was supported by the three scenario sensitivity analyses (data and analysis not shown) and was demonstrated considering the propensity score summarizing the effect of country, age, gender, body mass index, a history of venous thromboembolism, creatinine clearance, type of surgery, type of anesthesia, mobilization at discharge, and time between surgical closure and first fondaparinux injection, with an adjusted OR (95% confidence interval) of 0.70 (0.29 to 1.68) in favor of patients with a catheter. The rate of venous thromboembolism was comparable between patients in whom the catheter was removed between 1 and 2 days after surgery (4 of 665, 0.6%), and those in whom the catheter was removed between 3 and 6 days after surgery (9 of 863, 1.0%).

The overall rate of major bleeding was 0.8%, of which 71.4% (30 of 42) cases occurred within 15 days after surgery. Fatal bleeding was observed in five (0.1%) patients. Bleeding in a critical organ occurred in six (0.1%) patients. Bleeding at surgical site leading to reoperation occurred in 26 (0.5%) patients. Bleeding at nonsurgical site necessitating the transfusion of more than two units of blood occurred in five (0.1%) patients. There were no statistical differences in the rate of major bleeding between patients in whom the catheter was removed between 1 and 2 days after surgery (4 of 665, 0.6%), and those in whom the catheter was removed between 3 and 6 days after surgery (3 of 863, 0.3%). No neuraxial hematoma, perineural hematoma, or nerve damage was reported.

Twenty-three (0.4%) patients died 4 to 6 wk after surgery. Death was related to probable fatal pulmonary embolism or possible fatal pulmonary embolism in seven (0.1%) patients (Table 3).

Table 3: Symptomatic Venous Thromboembolism and Major Bleeds up to 5 ± 1 wk After Surgery (Intention-to-Treat Population)

DISCUSSION

The 1.0% overall rate of symptomatic venous thromboembolism after major orthopedic lower limb surgery in patients routinely treated with a once-daily subcutaneous injection of 2.5 mg fondaparinux, started 6 to 12 h after surgical closure and maintained for 4 ± 1 wk, compares favorably with the rates reported in trials or meta-analyses of extended thromboprophylaxis with low-molecular-weight heparins: 1.1%–2.3% after hip or knee replacement surgery (^13–18), and 0.8%–1.6% after hip fracture surgery (^19,20). In a meta-analysis of cohort studies of patients undergoing major orthopedic surgery with thromboprophylaxis, fatal pulmonary embolism was recorded in 0.18% of patients after elective hip or knee arthroplasty, and 0.30% after hip fracture surgery (²¹). In the EXPERT study, a lower rate of 0.13% was observed. Such data are new, because in the previous short-term randomized thromboprophylaxis trials, the incidence of symptomatic events during follow-up could not be interpreted because the natural history of the disease was modified by systematically performing bilateral venography 5 to 11 days after surgery and treating most patients with a positive venogram with therapeutic doses of anticoagulants (²²). Our results demonstrate that the efficacy of fondaparinux was maintained in patients in whom fondaparinux was withheld for 48 h to permit safe removal of a neuraxial or deep peripheral nerve catheter (⁵). Finally, the 1.0% overall incidence of major bleeding observed in the present study compares favorably with rates between 1.4% and 5.5% recorded in extended thromboprophylaxis clinical trials studying fondaparinux, enoxaparin, or aspirin in the same surgical settings using comparable criteria for defining major bleeding (^6,19,20,23).

The EXPERT study also demonstrated that neuraxial or deep peripheral nerve catheters can be used in patients receiving once-daily subcutaneous injections of 2.5 mg fondaparinux, provided that safety recommendations for catheter removal are followed. These recommendations were based on those established for low-molecular-weight heparins. Because the risk of hematoma is highest when the catheter is removed and the anticoagulant activity is high (⁹), a minimum of 10 to 12 h is recommended between the removal of the catheter and the last dose of low-molecular-weight heparin. This corresponds to a “drug free” period equal to about twice the half-life of the drug passed since the previous administration (²⁴). Similarly, because the half-life of fondaparinux is 17 h (¹¹), the drug free period was 36 h, and the administration of fondaparinux was reinitiated 48 h after the last injection. No neuraxial hematomas were reported in the 665 patients in whom the catheter was removed between 1 and 2 days after surgery, i.e., about 24 to 48 h after the last fondaparinux injection. Similarly, no patients experienced neuraxial hematomas among the 457 fondaparinux-treated patients after major abdominal surgery in whom the neuraxial catheter was withdrawn after the first postoperative injection of the study drug in the PEGASUS study. In that study, the administration of fondaparinux was reinitiated 25 h after the last injection (²⁵).

Although no epidural hematoma was reported in the 1553 patients with a neuraxial catheter in this study, this number may not be sufficient to draw a definitive conclusion regarding the safety of the concomitant use of fondaparinux and a neuraxial catheter. The rate of epidural hematoma in patients receiving neuraxial anesthesia is estimated to be between 1 in 150,000 and 1 in 220,000 (^26,27). This rate may be as high as 1 in 3000 in elderly patients treated with low-molecular-weight heparins for the prevention of venous thromboembolism (^26,28,29). Epidural hematoma is so rare that, in practice, it is very difficult to demonstrate the safety of an anticoagulant in patients undergoing neuraxial anesthesia (³⁰). The incidence of the risk of epidural hematoma after removal of a neuraxial catheter according to these guidelines is <0.24% (95% confidence interval upper bound). Likewise, the present study included only 78 patients with a deep peripheral nerve block catheter. The most we can conclude is that the incidence is <4.62% (95% confidence interval upper bound). The most important finding is that the efficacy of fondaparinux was not diminished when it was discontinued for 48 h to permit safe withdrawal of a neuraxial or deep peripheral nerve catheter. Several further sensitivity analyses confirmed the robustness of this result.

In conclusion, this prospective observational study confirmed the safety and efficacy of once-daily subcutaneous injection of 2.5 mg fondaparinux, given for 3 to 5 wk after major orthopedic lower limb surgery, which was previously reported in randomized clinical trials. Furthermore, once-daily subcutaneous injection of 2.5 mg fondaparinux seemed to be safe and effective in patients with a neuraxial or deep peripheral nerve catheter in whom the drug was discontinued for 48 h to permit safe removal of the catheter.

ACKNOWLEDGMENTS

The authors gratefully acknowledge Cécile Fragneaud (Lincoln, Boulogne-Billancourt, France) for her expertise in performing the statistical analysis.