Steven B. Deitelzweig, MD

Editor's Note: This letter concerns an article in a Cleveland Clinic Journal of Medicine supplement (Preventing Venous Thromboembolism Throughout the Continuum of Care) distributed to only a portion of the Journal's regular readership, owing to the terms of the grant supporting the supplement.

In Reply: We appreciate the comments by Drs. Fishmann and Boyd, but we strongly disagree with their suggestion that aspirin monotherapy is an appropriate option for the prevention of venous thromboembolism (VTE) following major orthopedic surgery.

As discussed in our original article,¹ multiple large-scale clinical trials in patients undergoing elective hip arthroplasty, knee arthroplasty, or hip fracture surgery have demonstrated the thromboprophylactic efficacy of warfarin, unfractionated heparin, low-molecular-weight heparin (LMWH), fondaparinux, and oral direct thrombin inhibitors. The relative risk reduction with these agents has been greater than 50% in most studies. In contrast, in a large meta-analysis of VTE prophylaxis following total hip replacement, which included data from 56 randomized trials published between 1966 and 1993, aspirin was not beneficial in preventing DVT.¹⁴

The largest prospective randomized trial comparing aspirin with placebo for VTE prevention was conducted between 1992 and 1998 among 17,444 patients in five countries.⁵ It involved 13,356 patients requiring hip fracture surgery and 4,088 patients requiring elective hip arthroplasty. Patients were randomized to receive aspirin 160 mg/day or placebo for 35 days. However, additional forms of VTE prophylaxis were allowed if deemed necessary by the clinician. In fact, 26% of patients received LMWH in addition to aspirin, and dual therapy was probably more common in those patients at highest thromboembolic risk. As such, the 36% relative risk reduction in VTE ascribed to aspirin should be viewed with caution. Further, this is a smaller risk reduction than that observed in trials of other anticoagulant agents.

A large, well-designed, randomized clinical trial comparing aspirin to LMWH or fondaparinux remains to be conducted.

Dr. Fishmann cites a small study of patients undergoing knee arthroplasty who received spinal anesthesia and intermittent calf compression devices.⁷ In this underpowered study, 275 patients were randomized to receive aspirin 325 mg twice daily or enoxaparin 30 mg twice daily for 3 weeks. The overall DVT rates were 14.1% in the enoxaparin group vs 17.8% in the aspirin group (P = .27).⁷ Patients who received aspirin had significantly more postoperative drainage than those randomized to enoxaparin. In addition, the protocol for scheduling enoxaparin 48 hours postoperatively is not consistent with recommendations of the American College of Chest Physicians (ACCP) and may have reduced the efficacy of enoxaparin.

The other evidence in support of aspirin cited by Dr. Fishmann includes an editorial,⁹ an uncontrolled retrospective analysis,⁸ a single-center retrospective review,¹⁰ and a review article.⁶ Although there is evidence that the use of aspirin is probably associated with a modest reduction in postoperative VTE risk, it has been unequivocally surpassed in efficacy by other anticoagulants.

Both the latest (2004) ACCP guidelines on VTE² and the 2006 International Consensus Statement on VTE prevention and treatment¹⁵ advise against aspirin monotherapy as VTE prophylaxis in any patient groups. It is likely that the upcoming 2008 ACCP guidelines will also advocate against using aspirin as well.

Lastly, the most recent guideline from the American Academy of Orthopaedic Surgeons advocating aspirin as monotherapy¹¹ is based on the assumption that the major important clinical end point in the orthopedic surgery patient is clinical pulmonary embolism, an end point that was not included as a lone primary end point in any of the modern randomized controlled studies in major orthopedic surgery. This represents a flawed logic for the development of evidence-based guideline recommendations, and this recommendation has not been advocated by well-respected bodies such as the ACCP and the international groups that developed the International Consensus Statement. Furthermore, if this practice is going to be advocated by the American Academy of Orthopaedic Surgeons, then large rigorously designed randomized trials must be conducted to compare aspirin to currently available anticoagulants, and the type of joint surgery should be clearly defined.

Editor's Note: This letter concerns an article in a Cleveland Clinic Journal of Medicine supplement (Preventing Venous Thromboembolism Throughout the Continuum of Care) distributed to only a portion of the Journal's regular readership, owing to the terms of the grant supporting the supplement.

In Reply: We appreciate the comments by Drs. Fishmann and Boyd, but we strongly disagree with their suggestion that aspirin monotherapy is an appropriate option for the prevention of venous thromboembolism (VTE) following major orthopedic surgery.

As discussed in our original article,¹ multiple large-scale clinical trials in patients undergoing elective hip arthroplasty, knee arthroplasty, or hip fracture surgery have demonstrated the thromboprophylactic efficacy of warfarin, unfractionated heparin, low-molecular-weight heparin (LMWH), fondaparinux, and oral direct thrombin inhibitors. The relative risk reduction with these agents has been greater than 50% in most studies. In contrast, in a large meta-analysis of VTE prophylaxis following total hip replacement, which included data from 56 randomized trials published between 1966 and 1993, aspirin was not beneficial in preventing DVT.¹⁴

The largest prospective randomized trial comparing aspirin with placebo for VTE prevention was conducted between 1992 and 1998 among 17,444 patients in five countries.⁵ It involved 13,356 patients requiring hip fracture surgery and 4,088 patients requiring elective hip arthroplasty. Patients were randomized to receive aspirin 160 mg/day or placebo for 35 days. However, additional forms of VTE prophylaxis were allowed if deemed necessary by the clinician. In fact, 26% of patients received LMWH in addition to aspirin, and dual therapy was probably more common in those patients at highest thromboembolic risk. As such, the 36% relative risk reduction in VTE ascribed to aspirin should be viewed with caution. Further, this is a smaller risk reduction than that observed in trials of other anticoagulant agents.

A large, well-designed, randomized clinical trial comparing aspirin to LMWH or fondaparinux remains to be conducted.

Dr. Fishmann cites a small study of patients undergoing knee arthroplasty who received spinal anesthesia and intermittent calf compression devices.⁷ In this underpowered study, 275 patients were randomized to receive aspirin 325 mg twice daily or enoxaparin 30 mg twice daily for 3 weeks. The overall DVT rates were 14.1% in the enoxaparin group vs 17.8% in the aspirin group (P = .27).⁷ Patients who received aspirin had significantly more postoperative drainage than those randomized to enoxaparin. In addition, the protocol for scheduling enoxaparin 48 hours postoperatively is not consistent with recommendations of the American College of Chest Physicians (ACCP) and may have reduced the efficacy of enoxaparin.

The other evidence in support of aspirin cited by Dr. Fishmann includes an editorial,⁹ an uncontrolled retrospective analysis,⁸ a single-center retrospective review,¹⁰ and a review article.⁶ Although there is evidence that the use of aspirin is probably associated with a modest reduction in postoperative VTE risk, it has been unequivocally surpassed in efficacy by other anticoagulants.

Both the latest (2004) ACCP guidelines on VTE² and the 2006 International Consensus Statement on VTE prevention and treatment¹⁵ advise against aspirin monotherapy as VTE prophylaxis in any patient groups. It is likely that the upcoming 2008 ACCP guidelines will also advocate against using aspirin as well.

Lastly, the most recent guideline from the American Academy of Orthopaedic Surgeons advocating aspirin as monotherapy¹¹ is based on the assumption that the major important clinical end point in the orthopedic surgery patient is clinical pulmonary embolism, an end point that was not included as a lone primary end point in any of the modern randomized controlled studies in major orthopedic surgery. This represents a flawed logic for the development of evidence-based guideline recommendations, and this recommendation has not been advocated by well-respected bodies such as the ACCP and the international groups that developed the International Consensus Statement. Furthermore, if this practice is going to be advocated by the American Academy of Orthopaedic Surgeons, then large rigorously designed randomized trials must be conducted to compare aspirin to currently available anticoagulants, and the type of joint surgery should be clearly defined.

Editor's Note: This letter concerns an article in a Cleveland Clinic Journal of Medicine supplement (Preventing Venous Thromboembolism Throughout the Continuum of Care) distributed to only a portion of the Journal's regular readership, owing to the terms of the grant supporting the supplement.

In Reply: We appreciate the comments by Drs. Fishmann and Boyd, but we strongly disagree with their suggestion that aspirin monotherapy is an appropriate option for the prevention of venous thromboembolism (VTE) following major orthopedic surgery.

As discussed in our original article,¹ multiple large-scale clinical trials in patients undergoing elective hip arthroplasty, knee arthroplasty, or hip fracture surgery have demonstrated the thromboprophylactic efficacy of warfarin, unfractionated heparin, low-molecular-weight heparin (LMWH), fondaparinux, and oral direct thrombin inhibitors. The relative risk reduction with these agents has been greater than 50% in most studies. In contrast, in a large meta-analysis of VTE prophylaxis following total hip replacement, which included data from 56 randomized trials published between 1966 and 1993, aspirin was not beneficial in preventing DVT.¹⁴

The largest prospective randomized trial comparing aspirin with placebo for VTE prevention was conducted between 1992 and 1998 among 17,444 patients in five countries.⁵ It involved 13,356 patients requiring hip fracture surgery and 4,088 patients requiring elective hip arthroplasty. Patients were randomized to receive aspirin 160 mg/day or placebo for 35 days. However, additional forms of VTE prophylaxis were allowed if deemed necessary by the clinician. In fact, 26% of patients received LMWH in addition to aspirin, and dual therapy was probably more common in those patients at highest thromboembolic risk. As such, the 36% relative risk reduction in VTE ascribed to aspirin should be viewed with caution. Further, this is a smaller risk reduction than that observed in trials of other anticoagulant agents.

A large, well-designed, randomized clinical trial comparing aspirin to LMWH or fondaparinux remains to be conducted.

Dr. Fishmann cites a small study of patients undergoing knee arthroplasty who received spinal anesthesia and intermittent calf compression devices.⁷ In this underpowered study, 275 patients were randomized to receive aspirin 325 mg twice daily or enoxaparin 30 mg twice daily for 3 weeks. The overall DVT rates were 14.1% in the enoxaparin group vs 17.8% in the aspirin group (P = .27).⁷ Patients who received aspirin had significantly more postoperative drainage than those randomized to enoxaparin. In addition, the protocol for scheduling enoxaparin 48 hours postoperatively is not consistent with recommendations of the American College of Chest Physicians (ACCP) and may have reduced the efficacy of enoxaparin.

The other evidence in support of aspirin cited by Dr. Fishmann includes an editorial,⁹ an uncontrolled retrospective analysis,⁸ a single-center retrospective review,¹⁰ and a review article.⁶ Although there is evidence that the use of aspirin is probably associated with a modest reduction in postoperative VTE risk, it has been unequivocally surpassed in efficacy by other anticoagulants.

Both the latest (2004) ACCP guidelines on VTE² and the 2006 International Consensus Statement on VTE prevention and treatment¹⁵ advise against aspirin monotherapy as VTE prophylaxis in any patient groups. It is likely that the upcoming 2008 ACCP guidelines will also advocate against using aspirin as well.

Lastly, the most recent guideline from the American Academy of Orthopaedic Surgeons advocating aspirin as monotherapy¹¹ is based on the assumption that the major important clinical end point in the orthopedic surgery patient is clinical pulmonary embolism, an end point that was not included as a lone primary end point in any of the modern randomized controlled studies in major orthopedic surgery. This represents a flawed logic for the development of evidence-based guideline recommendations, and this recommendation has not been advocated by well-respected bodies such as the ACCP and the international groups that developed the International Consensus Statement. Furthermore, if this practice is going to be advocated by the American Academy of Orthopaedic Surgeons, then large rigorously designed randomized trials must be conducted to compare aspirin to currently available anticoagulants, and the type of joint surgery should be clearly defined.

The need for prophylaxis of venous thromboembolism (VTE) in hospitalized acutely ill medical patients is well established. Without prophylaxis, hospitalized medical patients develop VTE at a rate of 5% to 15%.^1–3 Moreover, pulmonary embolism (PE) occurs more frequently in hospitalized medical patients than in nonmedical patients, and is a leading cause of sudden death in hospitalized medical patients.^4,5 Without appropriate prophylaxis, 1 in 20 hospitalized medical patients may suffer a fatal PE.⁴

PROPHYLAXIS IN MEDICAL PATIENTS: UNDERUSED AND OFTEN INAPPROPRIATE

Despite these risks and the clear indications for VTE prophylaxis in hospitalized medical patients, prophylaxis of VTE is omitted more often in these patients than in hospitalized surgical patients.⁵ Even when prophylaxis is given, it is often used inappropriately in the medical population. So concludes a recent analysis of data from 196,104 patients with acute medical conditions who were discharged from 227 US hospitals from January 2002 to September 2005.⁶ Criteria for inclusion in the analysis were patient age of 40 years or older, a hospital stay of 6 days or greater, and an absence of contraindications to anticoagulation. Appropriate prophylaxis was defined in accordance with the Sixth American College of Chest Physicians (ACCP) Consensus Conference on Antithrombotic Therapy.⁷

The analysis revealed an overall VTE prophylaxis rate of 61.8%, but the rate of appropriate prophylaxis was only 33.9%, meaning that two-thirds of discharged patients did not receive prophylaxis in accordance with ACCP guidelines. When temporal trends were analyzed according to groups based on patients’ diagnosis at admission (acute myocardial infarction, severe lung disease, ischemic stroke, cancer, heart failure, or trauma), the rate of appropriate prophylaxis remained essentially flat from the beginning to the end of the study period for virtually all diagnosis groups.⁶

Similar findings have emerged from the International Medical Prevention Registry on Venous Thromboembolism (IMPROVE), an ongoing international registry of acutely ill medical patients.⁸ Data from the first 15,156 patients, enrolled from July 2002 through September 2006, reveal that 50% of patients received pharmacologic and/or mechanical VTE prophylaxis in the hospital, and only 60% of patients who met established criteria for VTE prophylaxis actually received it.

Analysis of the US portion of the IMPROVE data shows that 54% of the US patient sample received some form of VTE prophylaxis; 22% of US patients received intermittent pneumatic compression, 21% received unfractionated heparin (UFH), 14% received low-molecular-weight heparin (LMWH), and 3% wore compression stockings.⁸ Thus, despite a paucity of data supporting a benefit of intermittent pneumatic compression in this population,⁹ it was the most frequently used form of prophylaxis in US patients.

CLINICAL TRIALS OF PHARMACOLOGIC PROPHYLAXIS IN MEDICAL PATIENTS

Reprinted, with permission, from New England Journal of Medicine (Francis CW. Prophylaxis for thromboembolism in hospitalized medical patients. N Engl J Med 2007; 356:1438–1444.). Copyright © 2007 Massachusetts Medical Society. All rights reserved.

The Prophylaxis in Medical Patients with Enoxaparin (MEDENOX) trial1 randomized 1,102 hospitalized patients to one of two doses of the LMWH enoxaparin (20 mg or 40 mg once daily subcutaneously) or placebo for 6 to 14 days. Compared with placebo, the 40-mg dose of enoxaparin was associated with a 63% reduction in risk of VTE over 3 months of follow-up (P < .001) (Figure 1).

The Prospective Evaluation of Dalteparin Efficacy for Prevention of VTE in Immobilized Patients Trial (PREVENT)2 was a multicenter, randomized, double-blind study comparing the LMWH dalteparin (5,000 IU daily given subcutaneously for 14 days) with placebo in 3,706 acutely ill medical patients. Over 90 days of follow-up, the risk of VTE was reduced by 44% in patients assigned to dalteparin compared with those assigned to placebo (P = .0015) (Figure 1).

The Arixtra for Thromboembolism Prevention in a Medical Indications Study (ARTEMIS)3 randomized 849 medical patients 60 years or older to 6 to 14 days of therapy with the selective factor Xa inhibitor fondaparinux (2.5 mg once daily subcutaneously) or placebo. Compared with the placebo group, fondaparinux recipients had a 47% lower risk of developing VTE by day 15 (P = .029) (Figure 1).

Fewer events and fatal PEs, but no effect on all-cause mortality

A recent meta-analysis by Dentali et al¹⁰ further demonstrates the efficacy of anticoagulant therapy for preventing symptomatic VTE in hospitalized medical patients. This analysis included several other trials in addition to the three reviewed above,^1–3 for a total of nine randomized studies (seven of which were dou-ble-blind) comprising 19,958 patients. Across the nine studies, anticoagulant prophylaxis was clearly superior to placebo in preventing fatal PE (relative risk, 0.38 [95% CI, 0.21 to 0.69]). There was a strong trend toward a reduction in symptomatic deep vein thrombosis (DVT) with prophylaxis but no effect on all-cause mortality. The meta-analysis also provided reassurance that prophylaxis does not increase the rate of major bleeding.

What the ACCP recommends

Current ACCP guidelines recommend the use of either LMWH or low-dose UFH (5,000 U subcutaneously two or three times daily) as a Grade 1A recommendation for VTE prophylaxis in patients with medical conditions and risk factors for VTE.⁹ This represents the guidelines’ highest level of recommendation, ie, one that is based on randomized controlled trials (RCTs) without important limitations. In contrast, the 2006 International Consensus Statement, developed as a collaborative effort among expert bodies on VTE, specified a more narrow dosing recommendation for UFH in this patient population (5,000 U three times daily, not twice daily) as well as specifying 40 mg once daily as the recommended dose of enoxaparin and 5,000 IU once daily as the recommended dose of dalteparin.¹¹

For medical patients with risk factors for VTE who have a contraindication to anticoagulant prophylaxis, the ACCP guidelines recommend the use of graduated compression stockings or intermittent pneumatic compression devices as a Grade 1C+ recommendation (“no RCTs but strong RCT results can be unequivocally extrapolated, or overwhelming evidence from observational studies”⁹).

Current ACCP guidelines do not address the use of fondaparinux in their recommendations for VTE prophylaxis in medical patients.

Getting a handle on bleeding risk

Patient characteristics that exclude pharmacologic thromboprophylaxis due to bleeding risk are generally limited to active bleeding or coagulopathy, as demonstrated by a platelet count less than 50,000 cells/µL or an international normalized ratio greater than 1.5. Additionally, bleeding risk should be carefully assessed if an invasive procedure is planned during a patient’s hospital stay.

It is worth noting that the anticoagulant doses used for VTE prophylaxis are a fraction of those used for treatment of VTE. Thus, if a patient would be treated with full-dose anticoagulation if VTE developed, then that patient should be eligible for VTE prophylaxis.

Because the use of mechanical forms of prophylaxis in medical patients is not truly evidence-based, mechanical prophylaxis should be reserved for medical patients who have a contraindication to anticoagulants, or for use in combination with anticoagulants in patients at very high risk of VTE.

UFH vs LMWH

Two meta-analyses have compared UFH with LMWH for VTE prevention in medical patients.12,13 In a recent analysis that included 10 trials directly comparing the two therapies, 14 trials comparing UFH with control, and 11 trials comparing LMWH with control, Wein et al found a lower risk of DVT with LMWH than with UFH (relative risk, 0.68 [95% CI, 0.52 to 0.88]) but no difference between the therapies in mortality or bleeding risk.¹² In an earlier and smaller analysis, Mismetti et al found no significant differences between UFH and LMWH in preventing DVT or death but did find a significant reduction in major bleeding episodes with LMWH versus three-times-daily UFH (52% relative reduction; P = .049).¹³

Randomized trials also reveal that enoxaparin 40 mg once daily is as efficacious as UFH 5,000 U three times daily for VTE prevention in medical patients.^14,15 The above analysis by Wein et al¹² and an additional meta-analysis by King and colleagues¹⁶ found that three-times-daily dosing of UFH is more efficacious than twice-daily dosing of UFH, but at the expense of more bleeding, including major bleeding.

Economic considerations

Because of differences in drug acquisition costs between UFH and the LMWH agents, several economic evaluations have compared the use of these therapies for prophylaxis in medical patients at risk of VTE.

In an analysis of hospital costs for medical patients receiving VTE prophylaxis from more than 330 US hospitals for the period 2001–2004, Burleigh et al found that mean total hospital costs were higher for patients who received UFH than for those who received LMWH ($7,615 vs $6,866) even though mean drug costs were higher for LMWH ($791 vs $569 for UFH).¹⁷ A reduction in hospital length of stay appeared to contribute to the overall savings with LMWH; other contributors may have included costs associated with heparin-induced thrombocytopenia (HIT) in UFH recipients or the extra nursing time required for administering UFH in two or three daily doses.

Leykum et al used a decision analysis model to estimate the economic effect of substituting enoxaparin for UFH in hospitalized medical patients for whom VTE prophylaxis is indicated.¹⁸ Cost data were based on Medicare reimbursement rates as well as drug and laboratory costs for a multi-institutional health system. The model assumed HIT incidence rates of 2.7% with UFH and 0.3% with enoxaparin. It also assumed the cost of a daily dose to be $4 for UFH versus $84 for enoxaparin. From the payer perspective, the model showed that substituting enoxaparin for UFH would reduce the overall cost of care by $28.61 per day on a per-patient basis, despite enoxaparin’s higher acquisition cost, and would save $4,550 per quality-adjusted life-year by reducing the incidence of HIT.

Another cost analysis confirms the association between HIT and increased hospital costs. Creekmore et al retrospectively analyzed data from 10,121 adult medical patients who received VTE prophylaxis at the University of Utah Hospital in Salt Lake City from August 2000 to November 2004.¹⁹ They found that an admission during which HIT developed incurred a mean cost of $56,364, compared with $15,231 for an admission without HIT. Because LMWH was associated with a lower incidence of HIT compared with UFH (0.084% vs 0.51%, respectively), LMWH reduced the incremental cost of VTE prophylaxis by $13.88 per patient compared with UFH.

THE EXCLAIM TRIAL: IS THERE A ROLE FOR EXTENDED PROPHYLAXIS?

Although the previously discussed studies have clearly demonstrated the benefit of in-hospital VTE prophylaxis for acutely ill medical patients, none has rigorously examined extended-duration out-of-hospital prophylaxis in these patients. This represents an important gap in the literature, since a substantial proportion of VTE develops in the outpatient setting within 3 months of a hospitalization, and most outpatient VTE episodes occur within 1 month of a preceding hospitalization.²⁰

To begin to fill this gap, the Extended Clinical Prophylaxis in Acutely Ill Medical Patients (EXCLAIM) trial was conducted to compare extended-duration LMWH prophylaxis with a standard LMWH prophylaxis regimen in acutely ill medical patients using a prospective, multicenter, randomized, double-blind, placebo-controlled design.²¹

Patients and study design

Patients were eligible for enrollment if they were aged 40 years or older and had recent immobilization (≤ 3 days), a predefined acute medical illness, and either level 1 mobility (total bed rest or sedentary state) or level 2 mobility (level 1 with bathroom privileges). The predefined acute medical illnesses consisted of New York Heart Association class III/IV heart failure, acute respiratory insufficiency, or other acute medical conditions, including post-acute ischemic stroke, acute infection without septic shock, and active cancer.

All patients received open-label enoxaparin 40 mg subcutaneously once daily for 10 ± 4 days, after which they were randomized to either enoxaparin 40 mg subcutaneously once daily or placebo for an additional 28 ± 4 days.

The primary efficacy end point was the incidence of VTE events, defined as asymptomatic DVT documented by mandatory ultrasonography at the end of the double-blind treatment period (28 ± 4 days) or as symptomatic DVT, symptomatic PE, or fatal PE at any time during the double-blind period. Symptomatic DVT was confirmed by objective tests; PE was confirmed by ventilation-perfusion scan, computed tomography, angiography, or autopsy. 

Secondary efficacy end points were mortality at the end of the double-blind period, at 3 months, and at 6 months, as well as the incidence of VTE at 3 months.

The primary safety outcome measure was the incidence of major hemorrhage during the double-blind period; secondary safety measures were rates of major and minor hemorrhage, minor hemorrhage, HIT, and serious adverse events.

After approximately half of the patients were enrolled, a planned and blinded interim analysis for futility concluded that the study was unlikely to show a statistically significant advantage of enoxaparin over placebo. The trial’s steering committee followed the suggestion of its data safety monitoring board to redefine the inclusion criteria to refocus enrollment on patients with a high risk of VTE. A blinded analysis was performed to identify this subgroup.

The resulting amended inclusion criteria were the same as above except that level 2 mobility had to be accompanied by at least one of three additional high-risk criteria: (1) age greater than 75 years, (2) history of prior VTE, or (3) diagnosis of cancer.

The trial’s main exclusion criteria were evidence of active bleeding, a contraindication to anticoagulation, receipt of prophylactic LMWH or UFH more than 72 hours prior to enrollment, treatment with an oral anticoagulant within 72 hours of enrollment, major surgery within the prior 3 months, cerebral stroke with bleeding, and persistent renal failure (creatinine clearance < 30 mL/min).

Results

The amended study population included 5,105 patients, 5,049 of whom received open-label enoxaparin. Of this group, 2,013 were randomized to active extended prophylaxis with enoxparain and 2,027 to placebo. Baseline characteristics, including level of mobility, were similar between the two groups.

The efficacy of extended prophylaxis with enoxaparin was enduring, as the cumulative incidence of VTE events at day 90 was significantly lower in enoxaparin recipients than in placebo recipients (3.0% vs 5.2%; relative reduction of 42%; P = .0115).

There was no difference in all-cause mortality at 6 months between the enoxaparin and placebo groups (10.1% vs 8.9%, respectively; P = .179).

Safety. Major hemorrhage was significantly more frequent in the enoxaparin arm, occurring in 0.60% of enoxaparin recipients compared with 0.15% of placebo recipients (P = .019). Minor bleeding was also more common with enoxaparin (5.20% vs 3.70%; P = .024).

Conclusions

The EXCLAIM trial found that an extended-duration (38-day) enoxaparin regimen significantly reduced the overall incidence of VTE relative to a 10-day enoxaparin regimen in acutely ill medical patients with reduced mobility. At the same time, the extended regimen was associated with a significant increase in the rate of major bleeding, although the incidence of major bleeding was low. The investigators concluded that the net clinical effect of extended-duration prophylaxis with enoxaparin is favorable, as only 46 patients would need to be treated to prevent one VTE event, whereas 224 patients would need to be treated to result in one major bleeding event.²¹

For this reason, it is reasonable to consider extended prophylaxis for hospitalized medical patients after identifying these patients’ risk factors. In keeping with the trial’s amended inclusion criteria, patients older than age 75 and those with cancer or prior VTE should receive special consideration for extended prophylaxis.

RECOMMENDED APPROACH TO VTE PREVENTION IN HOSPITALIZED MEDICAL PATIENTS

• All hospitalized medical patients should be screened at the time of admission, and patients at risk for VTE should receive prophylaxis.
• All patients with reduced mobility and one or more other risk factors for VTE are candidates for prophylaxis.
• Patients should be reassessed daily for the development of VTE risk factors during their hospitalization if risk factors are absent on admission.
• If screening or reassessment reveals any VTE risk factors, pharmacologic prophylaxis is the mainstay of therapy. If exclusion criteria for pharmacologic prophylaxis are present, mechanical prophylaxis with graduated compression stockings and intermittent compression devices should be used. For very high-risk medical patients without a contraindication to anticoagulants, combination prophylaxis with both an anticoagulant and mechanical devices is preferred.
• In this patient population, LMWH agents are preferred as pharmacologic prophylaxis over UFH and over fondaparinux (which is not currently approved by the US Food and Drug Administration for this population).
• If UFH is to be used in this patient population, 5,000 U three times daily is the preferred dosage.
• Extended pharmacologic prophylaxis should be considered in patients older than age 75 and in patients with a cancer diagnosis or a prior VTE episode.

Dr. Jaffer: Dr. Spyropoulos, are there any guidelines, other than those from the ACCP, that speak to VTE prophylaxis in hospitalized medical patients? If so, what are their take-home messages and how do they differ from the ACCP guidelines?

Dr. Spyropoulos: I was part of the group that developed the International Consensus Statement (ICS) published in International Angiology in 2006,¹¹ which is more recent than the latest ACCP guidelines, which were published in 2004. The ICS drew on much of the same data that the ACCP did, but we did an updated review of clinical trials.

For VTE prophylaxis in hospitalized medical patients, the ICS recommendations are more specific with regard to the type, dose, and dosing frequency of anticoagulant agents. First, they specify doses for both LMWH agents in this patient setting: 40 mg once daily for enoxaparin, and 5,000 IU once daily for dalteparin.

The ICS document also states that if UFH is the choice for prophylaxis, a regimen of 5,000 U three times daily should be considered. In the past year alone, two analyses suggest that three-times-daily dosing of UFH in medical patients provides superior efficacy to twice-daily dosing, although perhaps at the expense of more bleeding episodes.^12,16 It is important to remember that no large placebo-controlled trial supports the efficacy of a UFH regimen of 5,000 U twice daily in this population.

Finally, the ICS document states that fondaparinux 2.5 mg once daily is a viable option for prophylaxis in medical patients, based on the ARTEMIS trial,³ even though this represents an off-label use.

Dr. Jaffer: Real-world use of VTE prophylaxis is far from optimal, especially in medical patients, and this is partly a result of system-of-care issues. I’d like to conclude by asking each of my colleagues to offer your perspectives on how your own institutions have improved their systems of care to promote better use of VTE prophylaxis and what lessons might be shared with others. Dr. McKean, you work at Brigham and Women’s Hospital, which recently reported impressive results with an electronic alert system designed to increase clinicians’ consideration of VTE risk assessment and use of prophylaxis.²⁴ Please tell us about that study and the alert system.

Dr. McKean: Despite many educational initiatives at Brigham and Women’s Hospital, there were still some patients at high risk for VTE who were not receiving appropriate prophylaxis. What Dr. Samuel Goldhaber and his colleagues wanted to determine was whether changing the system of care could result in a reduced incidence of VTE.²⁴ They devised a computer software program linked to the patient database that used eight common risk factors to determine each hospitalized patient’s risk profile for VTE. Each risk factor was weighted according to a point scale, with major risk factors (cancer, prior VTE, or hypercoagulability) assigned 3 points, the intermediate risk factor of surgery assigned 2 points, and minor risk factors (advanced age, obesity, immobility, or use of hormone replacement therapy or oral contraceptives) assigned 1 point. For patients with a total risk score of 4 or greater, the computer screen generates a color-coded VTE risk alert that requires the physician to acknowledge the alert and choose one of three options: order prophylaxis as appropriate, review a 60-page document on the computer to learn more about prophylaxis, or do nothing.

The study found that hospitalized patients who were randomized to treatment under the computer alert system were significantly more likely to receive VTE prophylaxis and significantly less likely to develop VTE than were patients randomized to a control group. The alert system reduced the risk of DVT or PE at 90 days by 41% in patients considered to be at high risk. It was particularly interesting that the incidence of VTE was lower in the intervention group even when physicians chose not to use prophylaxis, which suggests that simply having this alert system in place improved outcomes, perhaps by raising awareness of the risk of VTE.²⁴

Additional studies are needed to better understand physicians’ behavior and determine why they seem to have a disproportionate fear of the risk of bleeding relative to the risk of clotting, including fatal PE, because that is really the heart of the matter. When patients are not given prophylaxis, often it is because of fear of bleeding. It is not clear, however, why some of these patients did not receive mechanical devices as an alternative form of prophylaxis, but this seems to be the case worldwide, as shown recently by the multinational ENDORSE study.²² Meanwhile, as we await studies to better understand physician perceptions and behaviors regarding prophylaxis, we need to work hard to change the system of care.

Dr. Deitelzweig: Over the past couple of years, the Ochsner Clinic has grown from a one-hospital teaching organization to a seven-hospital system with a mix of closed and open medical staff. The challenge is how to take a process that worked well in the one center, where appropriate prophylaxis was used about 90% of the time, and transfer it to the other centers in the larger system. We have endorsed several types of performance tools, such as the change-acceleration processes used by General Electric. The aim is to share a vision of heightening awareness. To do that, we have worked to mobilize the key stakeholders, at least half of them, to build algorithms that they all will endorse. It is easier said than done, however, and we have found it essential to involve both physicians and non-physician colleagues from pharmacy and nursing who have political and organizational clout.

Dr. Brotman: At Johns Hopkins, I took a bit more draconian approach to this issue because I thought that hospitalists often knew that they should be using VTE prophylaxis but sometimes weren’t, and I am not convinced that clinicians always look at prompts. So we came up with a system that incorporates both billing and documentation simultaneously. We put a hard stop on users’ documentation so that they could not sign off on a note or bill for their care until they checked off the kind of VTE prophylaxis they were using. Since hospitalists ultimately care about billing for their work, this system has at least ensured that everybody has considered and documented VTE prophylaxis on a daily basis. There are other hard stops that can be implemented in computer order-entry systems as well, and we are considering ways to roll them out on a broader scale.

However, all of these systems can have problems because patient situations change from day to day. For instance, VTE prophylaxis is not necessarily indicated in a 38-year-old ambulatory patient who comes in with a sickle cell crisis, but you will need to reconsider if the patient ends up in acute chest syndrome in the intensive care unit. I do not yet have a good way to ensure that this is being done on a daily basis with all patients.

Dr. Amin: At the University of California, Irvine, we implemented an electronic alert system, but we locked users in so that they could not complete their admission orders until they answered questions about VTE prevention. This practice increased our VTE prophylaxis rates tremendously. Because we are a level I trauma center, we allow users to bypass the screens one time, but the next time they log in, even to get a simple lab result, they have to answer the questions about VTE prevention.

With any system you develop, you also have to continue with the education process, because clinicians sometimes get into bad habits or simply forget things.

Dr. Spyropolous: At Lovelace Medical Center, we didn’t have the sophistication of an electronic order-entry system, but we had an experienced clinical pharmacist (the director of inpatient pharmacy) who helped to develop and champion VTE prevention guidelines that have then been used throughout the system in close conjunction with our hospitalists’ rounds. This system has been used successfully for the past 7 years.

The need for prophylaxis of venous thromboembolism (VTE) in hospitalized acutely ill medical patients is well established. Without prophylaxis, hospitalized medical patients develop VTE at a rate of 5% to 15%.^1–3 Moreover, pulmonary embolism (PE) occurs more frequently in hospitalized medical patients than in nonmedical patients, and is a leading cause of sudden death in hospitalized medical patients.^4,5 Without appropriate prophylaxis, 1 in 20 hospitalized medical patients may suffer a fatal PE.⁴

PROPHYLAXIS IN MEDICAL PATIENTS: UNDERUSED AND OFTEN INAPPROPRIATE

Despite these risks and the clear indications for VTE prophylaxis in hospitalized medical patients, prophylaxis of VTE is omitted more often in these patients than in hospitalized surgical patients.⁵ Even when prophylaxis is given, it is often used inappropriately in the medical population. So concludes a recent analysis of data from 196,104 patients with acute medical conditions who were discharged from 227 US hospitals from January 2002 to September 2005.⁶ Criteria for inclusion in the analysis were patient age of 40 years or older, a hospital stay of 6 days or greater, and an absence of contraindications to anticoagulation. Appropriate prophylaxis was defined in accordance with the Sixth American College of Chest Physicians (ACCP) Consensus Conference on Antithrombotic Therapy.⁷

The analysis revealed an overall VTE prophylaxis rate of 61.8%, but the rate of appropriate prophylaxis was only 33.9%, meaning that two-thirds of discharged patients did not receive prophylaxis in accordance with ACCP guidelines. When temporal trends were analyzed according to groups based on patients’ diagnosis at admission (acute myocardial infarction, severe lung disease, ischemic stroke, cancer, heart failure, or trauma), the rate of appropriate prophylaxis remained essentially flat from the beginning to the end of the study period for virtually all diagnosis groups.⁶

Similar findings have emerged from the International Medical Prevention Registry on Venous Thromboembolism (IMPROVE), an ongoing international registry of acutely ill medical patients.⁸ Data from the first 15,156 patients, enrolled from July 2002 through September 2006, reveal that 50% of patients received pharmacologic and/or mechanical VTE prophylaxis in the hospital, and only 60% of patients who met established criteria for VTE prophylaxis actually received it.

Analysis of the US portion of the IMPROVE data shows that 54% of the US patient sample received some form of VTE prophylaxis; 22% of US patients received intermittent pneumatic compression, 21% received unfractionated heparin (UFH), 14% received low-molecular-weight heparin (LMWH), and 3% wore compression stockings.⁸ Thus, despite a paucity of data supporting a benefit of intermittent pneumatic compression in this population,⁹ it was the most frequently used form of prophylaxis in US patients.

CLINICAL TRIALS OF PHARMACOLOGIC PROPHYLAXIS IN MEDICAL PATIENTS

Reprinted, with permission, from New England Journal of Medicine (Francis CW. Prophylaxis for thromboembolism in hospitalized medical patients. N Engl J Med 2007; 356:1438–1444.). Copyright © 2007 Massachusetts Medical Society. All rights reserved.

The Prophylaxis in Medical Patients with Enoxaparin (MEDENOX) trial1 randomized 1,102 hospitalized patients to one of two doses of the LMWH enoxaparin (20 mg or 40 mg once daily subcutaneously) or placebo for 6 to 14 days. Compared with placebo, the 40-mg dose of enoxaparin was associated with a 63% reduction in risk of VTE over 3 months of follow-up (P < .001) (Figure 1).

The Prospective Evaluation of Dalteparin Efficacy for Prevention of VTE in Immobilized Patients Trial (PREVENT)2 was a multicenter, randomized, double-blind study comparing the LMWH dalteparin (5,000 IU daily given subcutaneously for 14 days) with placebo in 3,706 acutely ill medical patients. Over 90 days of follow-up, the risk of VTE was reduced by 44% in patients assigned to dalteparin compared with those assigned to placebo (P = .0015) (Figure 1).

The Arixtra for Thromboembolism Prevention in a Medical Indications Study (ARTEMIS)3 randomized 849 medical patients 60 years or older to 6 to 14 days of therapy with the selective factor Xa inhibitor fondaparinux (2.5 mg once daily subcutaneously) or placebo. Compared with the placebo group, fondaparinux recipients had a 47% lower risk of developing VTE by day 15 (P = .029) (Figure 1).

Fewer events and fatal PEs, but no effect on all-cause mortality

A recent meta-analysis by Dentali et al¹⁰ further demonstrates the efficacy of anticoagulant therapy for preventing symptomatic VTE in hospitalized medical patients. This analysis included several other trials in addition to the three reviewed above,^1–3 for a total of nine randomized studies (seven of which were dou-ble-blind) comprising 19,958 patients. Across the nine studies, anticoagulant prophylaxis was clearly superior to placebo in preventing fatal PE (relative risk, 0.38 [95% CI, 0.21 to 0.69]). There was a strong trend toward a reduction in symptomatic deep vein thrombosis (DVT) with prophylaxis but no effect on all-cause mortality. The meta-analysis also provided reassurance that prophylaxis does not increase the rate of major bleeding.

What the ACCP recommends

Current ACCP guidelines recommend the use of either LMWH or low-dose UFH (5,000 U subcutaneously two or three times daily) as a Grade 1A recommendation for VTE prophylaxis in patients with medical conditions and risk factors for VTE.⁹ This represents the guidelines’ highest level of recommendation, ie, one that is based on randomized controlled trials (RCTs) without important limitations. In contrast, the 2006 International Consensus Statement, developed as a collaborative effort among expert bodies on VTE, specified a more narrow dosing recommendation for UFH in this patient population (5,000 U three times daily, not twice daily) as well as specifying 40 mg once daily as the recommended dose of enoxaparin and 5,000 IU once daily as the recommended dose of dalteparin.¹¹

For medical patients with risk factors for VTE who have a contraindication to anticoagulant prophylaxis, the ACCP guidelines recommend the use of graduated compression stockings or intermittent pneumatic compression devices as a Grade 1C+ recommendation (“no RCTs but strong RCT results can be unequivocally extrapolated, or overwhelming evidence from observational studies”⁹).

Current ACCP guidelines do not address the use of fondaparinux in their recommendations for VTE prophylaxis in medical patients.

Getting a handle on bleeding risk

Patient characteristics that exclude pharmacologic thromboprophylaxis due to bleeding risk are generally limited to active bleeding or coagulopathy, as demonstrated by a platelet count less than 50,000 cells/µL or an international normalized ratio greater than 1.5. Additionally, bleeding risk should be carefully assessed if an invasive procedure is planned during a patient’s hospital stay.

It is worth noting that the anticoagulant doses used for VTE prophylaxis are a fraction of those used for treatment of VTE. Thus, if a patient would be treated with full-dose anticoagulation if VTE developed, then that patient should be eligible for VTE prophylaxis.

Because the use of mechanical forms of prophylaxis in medical patients is not truly evidence-based, mechanical prophylaxis should be reserved for medical patients who have a contraindication to anticoagulants, or for use in combination with anticoagulants in patients at very high risk of VTE.

UFH vs LMWH

Two meta-analyses have compared UFH with LMWH for VTE prevention in medical patients.12,13 In a recent analysis that included 10 trials directly comparing the two therapies, 14 trials comparing UFH with control, and 11 trials comparing LMWH with control, Wein et al found a lower risk of DVT with LMWH than with UFH (relative risk, 0.68 [95% CI, 0.52 to 0.88]) but no difference between the therapies in mortality or bleeding risk.¹² In an earlier and smaller analysis, Mismetti et al found no significant differences between UFH and LMWH in preventing DVT or death but did find a significant reduction in major bleeding episodes with LMWH versus three-times-daily UFH (52% relative reduction; P = .049).¹³

Randomized trials also reveal that enoxaparin 40 mg once daily is as efficacious as UFH 5,000 U three times daily for VTE prevention in medical patients.^14,15 The above analysis by Wein et al¹² and an additional meta-analysis by King and colleagues¹⁶ found that three-times-daily dosing of UFH is more efficacious than twice-daily dosing of UFH, but at the expense of more bleeding, including major bleeding.

Economic considerations

Because of differences in drug acquisition costs between UFH and the LMWH agents, several economic evaluations have compared the use of these therapies for prophylaxis in medical patients at risk of VTE.

In an analysis of hospital costs for medical patients receiving VTE prophylaxis from more than 330 US hospitals for the period 2001–2004, Burleigh et al found that mean total hospital costs were higher for patients who received UFH than for those who received LMWH ($7,615 vs $6,866) even though mean drug costs were higher for LMWH ($791 vs $569 for UFH).¹⁷ A reduction in hospital length of stay appeared to contribute to the overall savings with LMWH; other contributors may have included costs associated with heparin-induced thrombocytopenia (HIT) in UFH recipients or the extra nursing time required for administering UFH in two or three daily doses.

Leykum et al used a decision analysis model to estimate the economic effect of substituting enoxaparin for UFH in hospitalized medical patients for whom VTE prophylaxis is indicated.¹⁸ Cost data were based on Medicare reimbursement rates as well as drug and laboratory costs for a multi-institutional health system. The model assumed HIT incidence rates of 2.7% with UFH and 0.3% with enoxaparin. It also assumed the cost of a daily dose to be $4 for UFH versus $84 for enoxaparin. From the payer perspective, the model showed that substituting enoxaparin for UFH would reduce the overall cost of care by $28.61 per day on a per-patient basis, despite enoxaparin’s higher acquisition cost, and would save $4,550 per quality-adjusted life-year by reducing the incidence of HIT.

Another cost analysis confirms the association between HIT and increased hospital costs. Creekmore et al retrospectively analyzed data from 10,121 adult medical patients who received VTE prophylaxis at the University of Utah Hospital in Salt Lake City from August 2000 to November 2004.¹⁹ They found that an admission during which HIT developed incurred a mean cost of $56,364, compared with $15,231 for an admission without HIT. Because LMWH was associated with a lower incidence of HIT compared with UFH (0.084% vs 0.51%, respectively), LMWH reduced the incremental cost of VTE prophylaxis by $13.88 per patient compared with UFH.

THE EXCLAIM TRIAL: IS THERE A ROLE FOR EXTENDED PROPHYLAXIS?

Although the previously discussed studies have clearly demonstrated the benefit of in-hospital VTE prophylaxis for acutely ill medical patients, none has rigorously examined extended-duration out-of-hospital prophylaxis in these patients. This represents an important gap in the literature, since a substantial proportion of VTE develops in the outpatient setting within 3 months of a hospitalization, and most outpatient VTE episodes occur within 1 month of a preceding hospitalization.²⁰

To begin to fill this gap, the Extended Clinical Prophylaxis in Acutely Ill Medical Patients (EXCLAIM) trial was conducted to compare extended-duration LMWH prophylaxis with a standard LMWH prophylaxis regimen in acutely ill medical patients using a prospective, multicenter, randomized, double-blind, placebo-controlled design.²¹

Patients and study design

Patients were eligible for enrollment if they were aged 40 years or older and had recent immobilization (≤ 3 days), a predefined acute medical illness, and either level 1 mobility (total bed rest or sedentary state) or level 2 mobility (level 1 with bathroom privileges). The predefined acute medical illnesses consisted of New York Heart Association class III/IV heart failure, acute respiratory insufficiency, or other acute medical conditions, including post-acute ischemic stroke, acute infection without septic shock, and active cancer.

All patients received open-label enoxaparin 40 mg subcutaneously once daily for 10 ± 4 days, after which they were randomized to either enoxaparin 40 mg subcutaneously once daily or placebo for an additional 28 ± 4 days.

The primary efficacy end point was the incidence of VTE events, defined as asymptomatic DVT documented by mandatory ultrasonography at the end of the double-blind treatment period (28 ± 4 days) or as symptomatic DVT, symptomatic PE, or fatal PE at any time during the double-blind period. Symptomatic DVT was confirmed by objective tests; PE was confirmed by ventilation-perfusion scan, computed tomography, angiography, or autopsy. 

Secondary efficacy end points were mortality at the end of the double-blind period, at 3 months, and at 6 months, as well as the incidence of VTE at 3 months.

The primary safety outcome measure was the incidence of major hemorrhage during the double-blind period; secondary safety measures were rates of major and minor hemorrhage, minor hemorrhage, HIT, and serious adverse events.

After approximately half of the patients were enrolled, a planned and blinded interim analysis for futility concluded that the study was unlikely to show a statistically significant advantage of enoxaparin over placebo. The trial’s steering committee followed the suggestion of its data safety monitoring board to redefine the inclusion criteria to refocus enrollment on patients with a high risk of VTE. A blinded analysis was performed to identify this subgroup.

The resulting amended inclusion criteria were the same as above except that level 2 mobility had to be accompanied by at least one of three additional high-risk criteria: (1) age greater than 75 years, (2) history of prior VTE, or (3) diagnosis of cancer.

The trial’s main exclusion criteria were evidence of active bleeding, a contraindication to anticoagulation, receipt of prophylactic LMWH or UFH more than 72 hours prior to enrollment, treatment with an oral anticoagulant within 72 hours of enrollment, major surgery within the prior 3 months, cerebral stroke with bleeding, and persistent renal failure (creatinine clearance < 30 mL/min).

Results

The amended study population included 5,105 patients, 5,049 of whom received open-label enoxaparin. Of this group, 2,013 were randomized to active extended prophylaxis with enoxparain and 2,027 to placebo. Baseline characteristics, including level of mobility, were similar between the two groups.

The efficacy of extended prophylaxis with enoxaparin was enduring, as the cumulative incidence of VTE events at day 90 was significantly lower in enoxaparin recipients than in placebo recipients (3.0% vs 5.2%; relative reduction of 42%; P = .0115).

There was no difference in all-cause mortality at 6 months between the enoxaparin and placebo groups (10.1% vs 8.9%, respectively; P = .179).

Safety. Major hemorrhage was significantly more frequent in the enoxaparin arm, occurring in 0.60% of enoxaparin recipients compared with 0.15% of placebo recipients (P = .019). Minor bleeding was also more common with enoxaparin (5.20% vs 3.70%; P = .024).

Conclusions

The EXCLAIM trial found that an extended-duration (38-day) enoxaparin regimen significantly reduced the overall incidence of VTE relative to a 10-day enoxaparin regimen in acutely ill medical patients with reduced mobility. At the same time, the extended regimen was associated with a significant increase in the rate of major bleeding, although the incidence of major bleeding was low. The investigators concluded that the net clinical effect of extended-duration prophylaxis with enoxaparin is favorable, as only 46 patients would need to be treated to prevent one VTE event, whereas 224 patients would need to be treated to result in one major bleeding event.²¹

For this reason, it is reasonable to consider extended prophylaxis for hospitalized medical patients after identifying these patients’ risk factors. In keeping with the trial’s amended inclusion criteria, patients older than age 75 and those with cancer or prior VTE should receive special consideration for extended prophylaxis.

RECOMMENDED APPROACH TO VTE PREVENTION IN HOSPITALIZED MEDICAL PATIENTS

• All hospitalized medical patients should be screened at the time of admission, and patients at risk for VTE should receive prophylaxis.
• All patients with reduced mobility and one or more other risk factors for VTE are candidates for prophylaxis.
• Patients should be reassessed daily for the development of VTE risk factors during their hospitalization if risk factors are absent on admission.
• If screening or reassessment reveals any VTE risk factors, pharmacologic prophylaxis is the mainstay of therapy. If exclusion criteria for pharmacologic prophylaxis are present, mechanical prophylaxis with graduated compression stockings and intermittent compression devices should be used. For very high-risk medical patients without a contraindication to anticoagulants, combination prophylaxis with both an anticoagulant and mechanical devices is preferred.
• In this patient population, LMWH agents are preferred as pharmacologic prophylaxis over UFH and over fondaparinux (which is not currently approved by the US Food and Drug Administration for this population).
• If UFH is to be used in this patient population, 5,000 U three times daily is the preferred dosage.
• Extended pharmacologic prophylaxis should be considered in patients older than age 75 and in patients with a cancer diagnosis or a prior VTE episode.

Dr. Jaffer: Dr. Spyropoulos, are there any guidelines, other than those from the ACCP, that speak to VTE prophylaxis in hospitalized medical patients? If so, what are their take-home messages and how do they differ from the ACCP guidelines?

Dr. Spyropoulos: I was part of the group that developed the International Consensus Statement (ICS) published in International Angiology in 2006,¹¹ which is more recent than the latest ACCP guidelines, which were published in 2004. The ICS drew on much of the same data that the ACCP did, but we did an updated review of clinical trials.

For VTE prophylaxis in hospitalized medical patients, the ICS recommendations are more specific with regard to the type, dose, and dosing frequency of anticoagulant agents. First, they specify doses for both LMWH agents in this patient setting: 40 mg once daily for enoxaparin, and 5,000 IU once daily for dalteparin.

The ICS document also states that if UFH is the choice for prophylaxis, a regimen of 5,000 U three times daily should be considered. In the past year alone, two analyses suggest that three-times-daily dosing of UFH in medical patients provides superior efficacy to twice-daily dosing, although perhaps at the expense of more bleeding episodes.^12,16 It is important to remember that no large placebo-controlled trial supports the efficacy of a UFH regimen of 5,000 U twice daily in this population.

Finally, the ICS document states that fondaparinux 2.5 mg once daily is a viable option for prophylaxis in medical patients, based on the ARTEMIS trial,³ even though this represents an off-label use.

Dr. Jaffer: Real-world use of VTE prophylaxis is far from optimal, especially in medical patients, and this is partly a result of system-of-care issues. I’d like to conclude by asking each of my colleagues to offer your perspectives on how your own institutions have improved their systems of care to promote better use of VTE prophylaxis and what lessons might be shared with others. Dr. McKean, you work at Brigham and Women’s Hospital, which recently reported impressive results with an electronic alert system designed to increase clinicians’ consideration of VTE risk assessment and use of prophylaxis.²⁴ Please tell us about that study and the alert system.

Dr. McKean: Despite many educational initiatives at Brigham and Women’s Hospital, there were still some patients at high risk for VTE who were not receiving appropriate prophylaxis. What Dr. Samuel Goldhaber and his colleagues wanted to determine was whether changing the system of care could result in a reduced incidence of VTE.²⁴ They devised a computer software program linked to the patient database that used eight common risk factors to determine each hospitalized patient’s risk profile for VTE. Each risk factor was weighted according to a point scale, with major risk factors (cancer, prior VTE, or hypercoagulability) assigned 3 points, the intermediate risk factor of surgery assigned 2 points, and minor risk factors (advanced age, obesity, immobility, or use of hormone replacement therapy or oral contraceptives) assigned 1 point. For patients with a total risk score of 4 or greater, the computer screen generates a color-coded VTE risk alert that requires the physician to acknowledge the alert and choose one of three options: order prophylaxis as appropriate, review a 60-page document on the computer to learn more about prophylaxis, or do nothing.

The study found that hospitalized patients who were randomized to treatment under the computer alert system were significantly more likely to receive VTE prophylaxis and significantly less likely to develop VTE than were patients randomized to a control group. The alert system reduced the risk of DVT or PE at 90 days by 41% in patients considered to be at high risk. It was particularly interesting that the incidence of VTE was lower in the intervention group even when physicians chose not to use prophylaxis, which suggests that simply having this alert system in place improved outcomes, perhaps by raising awareness of the risk of VTE.²⁴

Additional studies are needed to better understand physicians’ behavior and determine why they seem to have a disproportionate fear of the risk of bleeding relative to the risk of clotting, including fatal PE, because that is really the heart of the matter. When patients are not given prophylaxis, often it is because of fear of bleeding. It is not clear, however, why some of these patients did not receive mechanical devices as an alternative form of prophylaxis, but this seems to be the case worldwide, as shown recently by the multinational ENDORSE study.²² Meanwhile, as we await studies to better understand physician perceptions and behaviors regarding prophylaxis, we need to work hard to change the system of care.

Dr. Deitelzweig: Over the past couple of years, the Ochsner Clinic has grown from a one-hospital teaching organization to a seven-hospital system with a mix of closed and open medical staff. The challenge is how to take a process that worked well in the one center, where appropriate prophylaxis was used about 90% of the time, and transfer it to the other centers in the larger system. We have endorsed several types of performance tools, such as the change-acceleration processes used by General Electric. The aim is to share a vision of heightening awareness. To do that, we have worked to mobilize the key stakeholders, at least half of them, to build algorithms that they all will endorse. It is easier said than done, however, and we have found it essential to involve both physicians and non-physician colleagues from pharmacy and nursing who have political and organizational clout.

Dr. Brotman: At Johns Hopkins, I took a bit more draconian approach to this issue because I thought that hospitalists often knew that they should be using VTE prophylaxis but sometimes weren’t, and I am not convinced that clinicians always look at prompts. So we came up with a system that incorporates both billing and documentation simultaneously. We put a hard stop on users’ documentation so that they could not sign off on a note or bill for their care until they checked off the kind of VTE prophylaxis they were using. Since hospitalists ultimately care about billing for their work, this system has at least ensured that everybody has considered and documented VTE prophylaxis on a daily basis. There are other hard stops that can be implemented in computer order-entry systems as well, and we are considering ways to roll them out on a broader scale.

However, all of these systems can have problems because patient situations change from day to day. For instance, VTE prophylaxis is not necessarily indicated in a 38-year-old ambulatory patient who comes in with a sickle cell crisis, but you will need to reconsider if the patient ends up in acute chest syndrome in the intensive care unit. I do not yet have a good way to ensure that this is being done on a daily basis with all patients.

Dr. Amin: At the University of California, Irvine, we implemented an electronic alert system, but we locked users in so that they could not complete their admission orders until they answered questions about VTE prevention. This practice increased our VTE prophylaxis rates tremendously. Because we are a level I trauma center, we allow users to bypass the screens one time, but the next time they log in, even to get a simple lab result, they have to answer the questions about VTE prevention.

With any system you develop, you also have to continue with the education process, because clinicians sometimes get into bad habits or simply forget things.

Dr. Spyropolous: At Lovelace Medical Center, we didn’t have the sophistication of an electronic order-entry system, but we had an experienced clinical pharmacist (the director of inpatient pharmacy) who helped to develop and champion VTE prevention guidelines that have then been used throughout the system in close conjunction with our hospitalists’ rounds. This system has been used successfully for the past 7 years.

The need for prophylaxis of venous thromboembolism (VTE) in hospitalized acutely ill medical patients is well established. Without prophylaxis, hospitalized medical patients develop VTE at a rate of 5% to 15%.^1–3 Moreover, pulmonary embolism (PE) occurs more frequently in hospitalized medical patients than in nonmedical patients, and is a leading cause of sudden death in hospitalized medical patients.^4,5 Without appropriate prophylaxis, 1 in 20 hospitalized medical patients may suffer a fatal PE.⁴

PROPHYLAXIS IN MEDICAL PATIENTS: UNDERUSED AND OFTEN INAPPROPRIATE

Despite these risks and the clear indications for VTE prophylaxis in hospitalized medical patients, prophylaxis of VTE is omitted more often in these patients than in hospitalized surgical patients.⁵ Even when prophylaxis is given, it is often used inappropriately in the medical population. So concludes a recent analysis of data from 196,104 patients with acute medical conditions who were discharged from 227 US hospitals from January 2002 to September 2005.⁶ Criteria for inclusion in the analysis were patient age of 40 years or older, a hospital stay of 6 days or greater, and an absence of contraindications to anticoagulation. Appropriate prophylaxis was defined in accordance with the Sixth American College of Chest Physicians (ACCP) Consensus Conference on Antithrombotic Therapy.⁷

The analysis revealed an overall VTE prophylaxis rate of 61.8%, but the rate of appropriate prophylaxis was only 33.9%, meaning that two-thirds of discharged patients did not receive prophylaxis in accordance with ACCP guidelines. When temporal trends were analyzed according to groups based on patients’ diagnosis at admission (acute myocardial infarction, severe lung disease, ischemic stroke, cancer, heart failure, or trauma), the rate of appropriate prophylaxis remained essentially flat from the beginning to the end of the study period for virtually all diagnosis groups.⁶

Similar findings have emerged from the International Medical Prevention Registry on Venous Thromboembolism (IMPROVE), an ongoing international registry of acutely ill medical patients.⁸ Data from the first 15,156 patients, enrolled from July 2002 through September 2006, reveal that 50% of patients received pharmacologic and/or mechanical VTE prophylaxis in the hospital, and only 60% of patients who met established criteria for VTE prophylaxis actually received it.

Analysis of the US portion of the IMPROVE data shows that 54% of the US patient sample received some form of VTE prophylaxis; 22% of US patients received intermittent pneumatic compression, 21% received unfractionated heparin (UFH), 14% received low-molecular-weight heparin (LMWH), and 3% wore compression stockings.⁸ Thus, despite a paucity of data supporting a benefit of intermittent pneumatic compression in this population,⁹ it was the most frequently used form of prophylaxis in US patients.

CLINICAL TRIALS OF PHARMACOLOGIC PROPHYLAXIS IN MEDICAL PATIENTS

Reprinted, with permission, from New England Journal of Medicine (Francis CW. Prophylaxis for thromboembolism in hospitalized medical patients. N Engl J Med 2007; 356:1438–1444.). Copyright © 2007 Massachusetts Medical Society. All rights reserved.

The Prophylaxis in Medical Patients with Enoxaparin (MEDENOX) trial1 randomized 1,102 hospitalized patients to one of two doses of the LMWH enoxaparin (20 mg or 40 mg once daily subcutaneously) or placebo for 6 to 14 days. Compared with placebo, the 40-mg dose of enoxaparin was associated with a 63% reduction in risk of VTE over 3 months of follow-up (P < .001) (Figure 1).

The Prospective Evaluation of Dalteparin Efficacy for Prevention of VTE in Immobilized Patients Trial (PREVENT)2 was a multicenter, randomized, double-blind study comparing the LMWH dalteparin (5,000 IU daily given subcutaneously for 14 days) with placebo in 3,706 acutely ill medical patients. Over 90 days of follow-up, the risk of VTE was reduced by 44% in patients assigned to dalteparin compared with those assigned to placebo (P = .0015) (Figure 1).

The Arixtra for Thromboembolism Prevention in a Medical Indications Study (ARTEMIS)3 randomized 849 medical patients 60 years or older to 6 to 14 days of therapy with the selective factor Xa inhibitor fondaparinux (2.5 mg once daily subcutaneously) or placebo. Compared with the placebo group, fondaparinux recipients had a 47% lower risk of developing VTE by day 15 (P = .029) (Figure 1).

Fewer events and fatal PEs, but no effect on all-cause mortality

A recent meta-analysis by Dentali et al¹⁰ further demonstrates the efficacy of anticoagulant therapy for preventing symptomatic VTE in hospitalized medical patients. This analysis included several other trials in addition to the three reviewed above,^1–3 for a total of nine randomized studies (seven of which were dou-ble-blind) comprising 19,958 patients. Across the nine studies, anticoagulant prophylaxis was clearly superior to placebo in preventing fatal PE (relative risk, 0.38 [95% CI, 0.21 to 0.69]). There was a strong trend toward a reduction in symptomatic deep vein thrombosis (DVT) with prophylaxis but no effect on all-cause mortality. The meta-analysis also provided reassurance that prophylaxis does not increase the rate of major bleeding.

What the ACCP recommends

Current ACCP guidelines recommend the use of either LMWH or low-dose UFH (5,000 U subcutaneously two or three times daily) as a Grade 1A recommendation for VTE prophylaxis in patients with medical conditions and risk factors for VTE.⁹ This represents the guidelines’ highest level of recommendation, ie, one that is based on randomized controlled trials (RCTs) without important limitations. In contrast, the 2006 International Consensus Statement, developed as a collaborative effort among expert bodies on VTE, specified a more narrow dosing recommendation for UFH in this patient population (5,000 U three times daily, not twice daily) as well as specifying 40 mg once daily as the recommended dose of enoxaparin and 5,000 IU once daily as the recommended dose of dalteparin.¹¹

For medical patients with risk factors for VTE who have a contraindication to anticoagulant prophylaxis, the ACCP guidelines recommend the use of graduated compression stockings or intermittent pneumatic compression devices as a Grade 1C+ recommendation (“no RCTs but strong RCT results can be unequivocally extrapolated, or overwhelming evidence from observational studies”⁹).

Current ACCP guidelines do not address the use of fondaparinux in their recommendations for VTE prophylaxis in medical patients.

Getting a handle on bleeding risk

Patient characteristics that exclude pharmacologic thromboprophylaxis due to bleeding risk are generally limited to active bleeding or coagulopathy, as demonstrated by a platelet count less than 50,000 cells/µL or an international normalized ratio greater than 1.5. Additionally, bleeding risk should be carefully assessed if an invasive procedure is planned during a patient’s hospital stay.

It is worth noting that the anticoagulant doses used for VTE prophylaxis are a fraction of those used for treatment of VTE. Thus, if a patient would be treated with full-dose anticoagulation if VTE developed, then that patient should be eligible for VTE prophylaxis.

Because the use of mechanical forms of prophylaxis in medical patients is not truly evidence-based, mechanical prophylaxis should be reserved for medical patients who have a contraindication to anticoagulants, or for use in combination with anticoagulants in patients at very high risk of VTE.

UFH vs LMWH

Two meta-analyses have compared UFH with LMWH for VTE prevention in medical patients.12,13 In a recent analysis that included 10 trials directly comparing the two therapies, 14 trials comparing UFH with control, and 11 trials comparing LMWH with control, Wein et al found a lower risk of DVT with LMWH than with UFH (relative risk, 0.68 [95% CI, 0.52 to 0.88]) but no difference between the therapies in mortality or bleeding risk.¹² In an earlier and smaller analysis, Mismetti et al found no significant differences between UFH and LMWH in preventing DVT or death but did find a significant reduction in major bleeding episodes with LMWH versus three-times-daily UFH (52% relative reduction; P = .049).¹³

Randomized trials also reveal that enoxaparin 40 mg once daily is as efficacious as UFH 5,000 U three times daily for VTE prevention in medical patients.^14,15 The above analysis by Wein et al¹² and an additional meta-analysis by King and colleagues¹⁶ found that three-times-daily dosing of UFH is more efficacious than twice-daily dosing of UFH, but at the expense of more bleeding, including major bleeding.

Economic considerations

Because of differences in drug acquisition costs between UFH and the LMWH agents, several economic evaluations have compared the use of these therapies for prophylaxis in medical patients at risk of VTE.

In an analysis of hospital costs for medical patients receiving VTE prophylaxis from more than 330 US hospitals for the period 2001–2004, Burleigh et al found that mean total hospital costs were higher for patients who received UFH than for those who received LMWH ($7,615 vs $6,866) even though mean drug costs were higher for LMWH ($791 vs $569 for UFH).¹⁷ A reduction in hospital length of stay appeared to contribute to the overall savings with LMWH; other contributors may have included costs associated with heparin-induced thrombocytopenia (HIT) in UFH recipients or the extra nursing time required for administering UFH in two or three daily doses.

Leykum et al used a decision analysis model to estimate the economic effect of substituting enoxaparin for UFH in hospitalized medical patients for whom VTE prophylaxis is indicated.¹⁸ Cost data were based on Medicare reimbursement rates as well as drug and laboratory costs for a multi-institutional health system. The model assumed HIT incidence rates of 2.7% with UFH and 0.3% with enoxaparin. It also assumed the cost of a daily dose to be $4 for UFH versus $84 for enoxaparin. From the payer perspective, the model showed that substituting enoxaparin for UFH would reduce the overall cost of care by $28.61 per day on a per-patient basis, despite enoxaparin’s higher acquisition cost, and would save $4,550 per quality-adjusted life-year by reducing the incidence of HIT.

Another cost analysis confirms the association between HIT and increased hospital costs. Creekmore et al retrospectively analyzed data from 10,121 adult medical patients who received VTE prophylaxis at the University of Utah Hospital in Salt Lake City from August 2000 to November 2004.¹⁹ They found that an admission during which HIT developed incurred a mean cost of $56,364, compared with $15,231 for an admission without HIT. Because LMWH was associated with a lower incidence of HIT compared with UFH (0.084% vs 0.51%, respectively), LMWH reduced the incremental cost of VTE prophylaxis by $13.88 per patient compared with UFH.

THE EXCLAIM TRIAL: IS THERE A ROLE FOR EXTENDED PROPHYLAXIS?

Although the previously discussed studies have clearly demonstrated the benefit of in-hospital VTE prophylaxis for acutely ill medical patients, none has rigorously examined extended-duration out-of-hospital prophylaxis in these patients. This represents an important gap in the literature, since a substantial proportion of VTE develops in the outpatient setting within 3 months of a hospitalization, and most outpatient VTE episodes occur within 1 month of a preceding hospitalization.²⁰

To begin to fill this gap, the Extended Clinical Prophylaxis in Acutely Ill Medical Patients (EXCLAIM) trial was conducted to compare extended-duration LMWH prophylaxis with a standard LMWH prophylaxis regimen in acutely ill medical patients using a prospective, multicenter, randomized, double-blind, placebo-controlled design.²¹

Patients and study design

Patients were eligible for enrollment if they were aged 40 years or older and had recent immobilization (≤ 3 days), a predefined acute medical illness, and either level 1 mobility (total bed rest or sedentary state) or level 2 mobility (level 1 with bathroom privileges). The predefined acute medical illnesses consisted of New York Heart Association class III/IV heart failure, acute respiratory insufficiency, or other acute medical conditions, including post-acute ischemic stroke, acute infection without septic shock, and active cancer.

All patients received open-label enoxaparin 40 mg subcutaneously once daily for 10 ± 4 days, after which they were randomized to either enoxaparin 40 mg subcutaneously once daily or placebo for an additional 28 ± 4 days.

The primary efficacy end point was the incidence of VTE events, defined as asymptomatic DVT documented by mandatory ultrasonography at the end of the double-blind treatment period (28 ± 4 days) or as symptomatic DVT, symptomatic PE, or fatal PE at any time during the double-blind period. Symptomatic DVT was confirmed by objective tests; PE was confirmed by ventilation-perfusion scan, computed tomography, angiography, or autopsy. 

Secondary efficacy end points were mortality at the end of the double-blind period, at 3 months, and at 6 months, as well as the incidence of VTE at 3 months.

The primary safety outcome measure was the incidence of major hemorrhage during the double-blind period; secondary safety measures were rates of major and minor hemorrhage, minor hemorrhage, HIT, and serious adverse events.

After approximately half of the patients were enrolled, a planned and blinded interim analysis for futility concluded that the study was unlikely to show a statistically significant advantage of enoxaparin over placebo. The trial’s steering committee followed the suggestion of its data safety monitoring board to redefine the inclusion criteria to refocus enrollment on patients with a high risk of VTE. A blinded analysis was performed to identify this subgroup.

The resulting amended inclusion criteria were the same as above except that level 2 mobility had to be accompanied by at least one of three additional high-risk criteria: (1) age greater than 75 years, (2) history of prior VTE, or (3) diagnosis of cancer.

The trial’s main exclusion criteria were evidence of active bleeding, a contraindication to anticoagulation, receipt of prophylactic LMWH or UFH more than 72 hours prior to enrollment, treatment with an oral anticoagulant within 72 hours of enrollment, major surgery within the prior 3 months, cerebral stroke with bleeding, and persistent renal failure (creatinine clearance < 30 mL/min).

Results

The amended study population included 5,105 patients, 5,049 of whom received open-label enoxaparin. Of this group, 2,013 were randomized to active extended prophylaxis with enoxparain and 2,027 to placebo. Baseline characteristics, including level of mobility, were similar between the two groups.

The efficacy of extended prophylaxis with enoxaparin was enduring, as the cumulative incidence of VTE events at day 90 was significantly lower in enoxaparin recipients than in placebo recipients (3.0% vs 5.2%; relative reduction of 42%; P = .0115).

There was no difference in all-cause mortality at 6 months between the enoxaparin and placebo groups (10.1% vs 8.9%, respectively; P = .179).

Safety. Major hemorrhage was significantly more frequent in the enoxaparin arm, occurring in 0.60% of enoxaparin recipients compared with 0.15% of placebo recipients (P = .019). Minor bleeding was also more common with enoxaparin (5.20% vs 3.70%; P = .024).

Conclusions

The EXCLAIM trial found that an extended-duration (38-day) enoxaparin regimen significantly reduced the overall incidence of VTE relative to a 10-day enoxaparin regimen in acutely ill medical patients with reduced mobility. At the same time, the extended regimen was associated with a significant increase in the rate of major bleeding, although the incidence of major bleeding was low. The investigators concluded that the net clinical effect of extended-duration prophylaxis with enoxaparin is favorable, as only 46 patients would need to be treated to prevent one VTE event, whereas 224 patients would need to be treated to result in one major bleeding event.²¹

For this reason, it is reasonable to consider extended prophylaxis for hospitalized medical patients after identifying these patients’ risk factors. In keeping with the trial’s amended inclusion criteria, patients older than age 75 and those with cancer or prior VTE should receive special consideration for extended prophylaxis.

RECOMMENDED APPROACH TO VTE PREVENTION IN HOSPITALIZED MEDICAL PATIENTS

• All hospitalized medical patients should be screened at the time of admission, and patients at risk for VTE should receive prophylaxis.
• All patients with reduced mobility and one or more other risk factors for VTE are candidates for prophylaxis.
• Patients should be reassessed daily for the development of VTE risk factors during their hospitalization if risk factors are absent on admission.
• If screening or reassessment reveals any VTE risk factors, pharmacologic prophylaxis is the mainstay of therapy. If exclusion criteria for pharmacologic prophylaxis are present, mechanical prophylaxis with graduated compression stockings and intermittent compression devices should be used. For very high-risk medical patients without a contraindication to anticoagulants, combination prophylaxis with both an anticoagulant and mechanical devices is preferred.
• In this patient population, LMWH agents are preferred as pharmacologic prophylaxis over UFH and over fondaparinux (which is not currently approved by the US Food and Drug Administration for this population).
• If UFH is to be used in this patient population, 5,000 U three times daily is the preferred dosage.
• Extended pharmacologic prophylaxis should be considered in patients older than age 75 and in patients with a cancer diagnosis or a prior VTE episode.

Dr. Jaffer: Dr. Spyropoulos, are there any guidelines, other than those from the ACCP, that speak to VTE prophylaxis in hospitalized medical patients? If so, what are their take-home messages and how do they differ from the ACCP guidelines?

Dr. Spyropoulos: I was part of the group that developed the International Consensus Statement (ICS) published in International Angiology in 2006,¹¹ which is more recent than the latest ACCP guidelines, which were published in 2004. The ICS drew on much of the same data that the ACCP did, but we did an updated review of clinical trials.

For VTE prophylaxis in hospitalized medical patients, the ICS recommendations are more specific with regard to the type, dose, and dosing frequency of anticoagulant agents. First, they specify doses for both LMWH agents in this patient setting: 40 mg once daily for enoxaparin, and 5,000 IU once daily for dalteparin.

The ICS document also states that if UFH is the choice for prophylaxis, a regimen of 5,000 U three times daily should be considered. In the past year alone, two analyses suggest that three-times-daily dosing of UFH in medical patients provides superior efficacy to twice-daily dosing, although perhaps at the expense of more bleeding episodes.^12,16 It is important to remember that no large placebo-controlled trial supports the efficacy of a UFH regimen of 5,000 U twice daily in this population.

Finally, the ICS document states that fondaparinux 2.5 mg once daily is a viable option for prophylaxis in medical patients, based on the ARTEMIS trial,³ even though this represents an off-label use.

Dr. Jaffer: Real-world use of VTE prophylaxis is far from optimal, especially in medical patients, and this is partly a result of system-of-care issues. I’d like to conclude by asking each of my colleagues to offer your perspectives on how your own institutions have improved their systems of care to promote better use of VTE prophylaxis and what lessons might be shared with others. Dr. McKean, you work at Brigham and Women’s Hospital, which recently reported impressive results with an electronic alert system designed to increase clinicians’ consideration of VTE risk assessment and use of prophylaxis.²⁴ Please tell us about that study and the alert system.

Dr. McKean: Despite many educational initiatives at Brigham and Women’s Hospital, there were still some patients at high risk for VTE who were not receiving appropriate prophylaxis. What Dr. Samuel Goldhaber and his colleagues wanted to determine was whether changing the system of care could result in a reduced incidence of VTE.²⁴ They devised a computer software program linked to the patient database that used eight common risk factors to determine each hospitalized patient’s risk profile for VTE. Each risk factor was weighted according to a point scale, with major risk factors (cancer, prior VTE, or hypercoagulability) assigned 3 points, the intermediate risk factor of surgery assigned 2 points, and minor risk factors (advanced age, obesity, immobility, or use of hormone replacement therapy or oral contraceptives) assigned 1 point. For patients with a total risk score of 4 or greater, the computer screen generates a color-coded VTE risk alert that requires the physician to acknowledge the alert and choose one of three options: order prophylaxis as appropriate, review a 60-page document on the computer to learn more about prophylaxis, or do nothing.

The study found that hospitalized patients who were randomized to treatment under the computer alert system were significantly more likely to receive VTE prophylaxis and significantly less likely to develop VTE than were patients randomized to a control group. The alert system reduced the risk of DVT or PE at 90 days by 41% in patients considered to be at high risk. It was particularly interesting that the incidence of VTE was lower in the intervention group even when physicians chose not to use prophylaxis, which suggests that simply having this alert system in place improved outcomes, perhaps by raising awareness of the risk of VTE.²⁴

Additional studies are needed to better understand physicians’ behavior and determine why they seem to have a disproportionate fear of the risk of bleeding relative to the risk of clotting, including fatal PE, because that is really the heart of the matter. When patients are not given prophylaxis, often it is because of fear of bleeding. It is not clear, however, why some of these patients did not receive mechanical devices as an alternative form of prophylaxis, but this seems to be the case worldwide, as shown recently by the multinational ENDORSE study.²² Meanwhile, as we await studies to better understand physician perceptions and behaviors regarding prophylaxis, we need to work hard to change the system of care.

Dr. Deitelzweig: Over the past couple of years, the Ochsner Clinic has grown from a one-hospital teaching organization to a seven-hospital system with a mix of closed and open medical staff. The challenge is how to take a process that worked well in the one center, where appropriate prophylaxis was used about 90% of the time, and transfer it to the other centers in the larger system. We have endorsed several types of performance tools, such as the change-acceleration processes used by General Electric. The aim is to share a vision of heightening awareness. To do that, we have worked to mobilize the key stakeholders, at least half of them, to build algorithms that they all will endorse. It is easier said than done, however, and we have found it essential to involve both physicians and non-physician colleagues from pharmacy and nursing who have political and organizational clout.

Dr. Brotman: At Johns Hopkins, I took a bit more draconian approach to this issue because I thought that hospitalists often knew that they should be using VTE prophylaxis but sometimes weren’t, and I am not convinced that clinicians always look at prompts. So we came up with a system that incorporates both billing and documentation simultaneously. We put a hard stop on users’ documentation so that they could not sign off on a note or bill for their care until they checked off the kind of VTE prophylaxis they were using. Since hospitalists ultimately care about billing for their work, this system has at least ensured that everybody has considered and documented VTE prophylaxis on a daily basis. There are other hard stops that can be implemented in computer order-entry systems as well, and we are considering ways to roll them out on a broader scale.

However, all of these systems can have problems because patient situations change from day to day. For instance, VTE prophylaxis is not necessarily indicated in a 38-year-old ambulatory patient who comes in with a sickle cell crisis, but you will need to reconsider if the patient ends up in acute chest syndrome in the intensive care unit. I do not yet have a good way to ensure that this is being done on a daily basis with all patients.

Dr. Amin: At the University of California, Irvine, we implemented an electronic alert system, but we locked users in so that they could not complete their admission orders until they answered questions about VTE prevention. This practice increased our VTE prophylaxis rates tremendously. Because we are a level I trauma center, we allow users to bypass the screens one time, but the next time they log in, even to get a simple lab result, they have to answer the questions about VTE prevention.

With any system you develop, you also have to continue with the education process, because clinicians sometimes get into bad habits or simply forget things.

Dr. Spyropolous: At Lovelace Medical Center, we didn’t have the sophistication of an electronic order-entry system, but we had an experienced clinical pharmacist (the director of inpatient pharmacy) who helped to develop and champion VTE prevention guidelines that have then been used throughout the system in close conjunction with our hospitalists’ rounds. This system has been used successfully for the past 7 years.

Venous thromboembolism (VTE) is a major complication of cancer, occurring in 4% to 20% of patients,¹ and is one of the leading causes of death in cancer patients, although these figures are believed to be underestimates, given the low autopsy rates among cancer patients.² In hospitalized cancer patients specifically, VTE is the second leading cause of death.^3,4 The risk of VTE in cancer patients undergoing surgery is three to five times greater than that in surgical patients without cancer.⁴ Moreover, cancer patients with symptomatic deep vein thrombosis (DVT) exhibit a high risk of recurrent VTE that may persist for many years after the index event.⁵

VTE PREVENTION POSES PARTICULAR CHALLENGES IN CANCER PATIENTS

Until recently, data on VTE prevention specific to cancer patients have been sparse. Cancer patients have represented only a small subset (< 20%) of participants in most of the largest clinical trials of VTE prophylaxis. Until the past 2 or 3 years, clinicians largely have had to extrapolate their approach to VTE prophylaxis in cancer patients from data in patients without cancer, bearing in mind that cancer patients are among the populations at highest risk of developing VTE.

High rates of VTE, even with prophylaxis

Further insights have come from the @RISTOS project, a Web-based prospective registry of patients undergoing general, urologic, or gynecologic surgery for cancer at multiple centers in Italy.⁸ Of the 2,372 patients tracked in this study, 82% received in-hospital VTE prophylaxis and 31% received prophylaxis following discharge. Despite this relatively high frequency of prophylaxis, however, the incidence of clinically overt VTE was 2.1% and the incidence of fatal VTE was 0.8%. Notably, most VTE events occurred after hospital discharge, and VTE was the most common cause of 30-day postoperative death in this registry.

RISK FACTORS: CANCER TYPE AND TREATMENT LOOM LARGE

Both the type and stage of a patient’s cancer are important in assessing the risk of VTE. For men, cancers of the prostate, colon, brain, and lung have been associated with an increased risk of VTE. Among women, cancers of the breast, ovary, and lung have been especially implicated as risk factors for VTE.^9,10

The type of cancer therapy also influences VTE risk:

• Surgery. Among patients who undergo cancer-related surgery, the rate of proximal DVT is 10% to 20%, the rate of clinically evident PE is 4% to 10%, and the incidence of fatal PE is 0.2% to 5%.^8,11
• Systemic treatments, including chemotherapy and hormone therapy, are also associated with an increased risk of VTE.^12–15
• Central venous catheters. Approximately 4% of cancer patients who have central venous catheters placed develop clinically relevant VTE.^16,17

In addition to the above risks related to cancer treatments, the following have been identified as risk factors for VTE in surgical oncology patients:

• Age greater than 40 years (risk also increases steeply after age 60 and again after age 75)
• Cancer procoagulants
• Thrombophilia
• Length and complications of cancer surgery (ie, often involving tissue trauma and immobilization)
• Debilitation and slow recovery.

Another risk factor worth noting is perioperative transfusion, as illustrated in a recent study of 14,104 adults undergoing colorectal cancer resection.¹⁸ The overall incidence of VTE in these patients was 1.0%, and the risk of death was nearly four times as great in patients who developed VTE as in those who did not. Notably, the need for transfusion was a marker of increased risk of VTE, particularly in women: women who received perioperative transfusions had almost double the risk of developing VTE compared with women who did not receive transfusions (P = .004).

CLINICAL TRIALS OF PROPHYLAXIS IN CANCER SURGERY PATIENTS

LMWH vs UFH for in-hospital prophylaxis

Two large randomized, double-blind trials have compared low-molecular-weight heparin (LMWH) with low-dose unfractionated heparin (UFH) for VTE prophylaxis in surgical patients with cancer—the Enoxaparin and Cancer (ENOXACAN) study¹⁹ and the Canadian Colorectal Surgery DVT Prophylaxis Trial.²⁰ Patients in these studies underwent surgery for abdominal or pelvic cancer (mostly colorectal cancer). Both studies compared 40 mg of the LMWH enoxaparin given once daily with 5,000 U of UFH given three times daily for 7 to 10 days postoperatively. Outcome measures were the presence of DVT determined by venography on day 7 to 10 and the incidence of symptomatic VTE. Rates of VTE were statistically equivalent between the two treatment arms in both ENOXACAN (14.7% with LMWH vs 18.2% with UFH) and the Canadian Colorectal Surgery study (9.4% with both therapies), as were rates of major bleeding (4.1% with LMWH vs 2.9% with UFH in ENOXACAN; 2.7% with LMWH vs 1.5% with UFH in the Canadian study).

These findings are consistent with a 2001 meta-analysis by Mismetti et al of all available randomized trials comparing LMWH with placebo or with UFH for VTE prophylaxis in general surgery.²¹ This analysis found no differences in rates of asymptomatic DVT, clinical PE, clinical thromboembolism, death, major hemorrhage, total hemorrhage, wound hematoma, or need for transfusion between LMWH and UFH in patients undergoing either cancer-related surgery or surgery not related to cancer.

Fondaparinux for in-hospital prophylaxis

Subgroup analysis of the large randomized trial known as PEGASUS²² sheds some light on the efficacy of the factor Xa inhibitor fondaparinux relative to LMWH for thromboprophylaxis in cancer surgery patients. PEGASUS compared fondaparinux 2.5 mg once daily with the LMWH dalteparin 5,000 IU once daily for 5 to 9 days in patients undergoing high-risk abdominal surgery. Among the study’s 1,408 patients undergoing surgery for cancer, rates of VTE were 4.7% in the fondaparinux group compared with 7.7% in the LMWH group, a relative risk reduction of 38.6% with fondaparinux (95% CI, 6.7% to 59.6%). In contrast, in the rest of the PEGASUS population (patients undergoing abdominal surgery for reasons other than cancer), LMWH was nonsignificantly more efficacious at preventing VTE than was fondaparinux. Rates of major bleeding in this cancer subgroup were comparable between the two treatments.

Two additional randomized trials have evaluated extended prophylaxis with LMWH in surgical cancer patients—ENOXACAN II²³ and the Fragmin After Major Abdominal Surgery (FAME) study.²⁴

In ENOXACAN II, patients undergoing surgery for abdominal or pelvic cancer first received 6 to 10 days of prophylaxis with enoxaparin 40 mg once daily and then were randomized in a double-blind fashion to an additional 21 days of enoxaparin or placebo.²³ Among 332 patients in the intent-to-treat analysis, the rate of VTE at the end of the double-blind phase was reduced from 12.0% with placebo to 4.8% with extended-duration enoxaparin (P = .02), an effect that was maintained at 3-month follow-up (P = .01). There was no significant difference between the two groups in rates of major bleeding events or any bleeding events.

In FAME, patients received 5,000 IU of dalteparin once daily for 1 week following major abdominal surgery and then were randomized in open-label fashion to either placebo or extended prophylaxis with dalteparin for 3 more weeks; a subanalysis examined outcomes in the 198 FAME participants whose abdominal surgery was for cancer.²⁴ Among these 198 cancer surgery patients, the rate of venography-documented VTE at 4 weeks was reduced from 19.6% with placebo to 8.8% with extended-duration dalteparin, a relative reduction of 55% (P = .03). The rate of proximal DVT was reduced from 10.4% to 2.2% with extended prophylaxis, a relative reduction of 79% (P = .02).

The number needed to treat with extended LMWH prophylaxis to prevent one VTE event was 14 in ENOXACAN II²³ and 9 in the FAME subanalysis of cancer surgery patients.²⁴

New systematic review of relevant trials

Leonardi et al recently published a systematic review of 26 randomized controlled trials of DVT prophylaxis in 7,639 cancer surgery patients.²⁵ They found the overall incidence of DVT to be 12.7% in those who received pharmacologic prophylaxis compared with 35.2% in controls. They also found high-dose LMWH therapy (> 3,400 U daily) to be associated with a significantly lower incidence of DVT than low-dose LMWH therapy (≤ 3,400 U daily) (7.9% vs 14.5%, respectively; P < .01). No differences were demonstrated between LMWH and UFH in preventing DVT, DVT location, or bleeding. Bleeding complications requiring discontinuation of pharmacologic prophylaxis occurred in 3% of patients overall.

Implications of HIT

The sequelae of heparin-induced thrombocytopenia (HIT) can have major consequences for cancer surgery patients. The incidence of HIT is markedly lower with LMWH than with UFH, as demonstrated in a nested case-control study by Creekmore et al.²⁶ These researchers also found that the average cost of an admission during which HIT developed was nearly four times as great as the average cost of an admission during which UFH or LMWH was given without development of HIT ($56,364 vs $15,231; P < .001).

EVIDENCE IN SPECIFIC ONCOLOGIC POPULATIONS

Most of the patients in the trials reviewed above underwent abdominal surgery for malignancy. Although studies of VTE prophylaxis in patients undergoing nonabdominal cancer surgery are relatively few, some data are available for a few other specific oncologic populations, as reviewed below.

Surgery for gynecologic cancer

There is a paucity of randomized controlled trials or prospective observational studies on VTE and its prevention in the gynecologic cancer surgery population. Based on small historical studies, the postoperative risk of VTE in this population varies from 12% to 35%.^27,28 Twice-daily administration of UFH 5,000 U appears to be ineffective as VTE prophylaxis in this population, but increasing the frequency to three times daily reduces VTE risk by 50% to 60% compared with placebo. Once-daily LMWH is comparable to three-times-daily UFH in efficacy and safety in this population.

A systematic Cochrane review of eight randomized controlled trials in patients undergoing major gynecologic surgery revealed that heparin prophylaxis (either UFH or LMWH) reduces the risk of DVT by 70% compared with no prophylaxis, with an identical risk reduction specifically among women with malignancy (odds ratio, 0.30; 95% CI, 0.10 to 0.89).²⁹ This review found no evidence that anticoagulation reduces the risk of PE following major gynecologic surgery. LMWH and UFH were similar in efficacy for preventing DVT and had a comparable risk of bleeding complications.

Surgery for urologic cancer

The risk of VTE and the benefits of thromboprophylaxis also are poorly studied in patients undergoing surgery for urologic cancer.

The risk of VTE varies with the type of urologic surgery and the method used to diagnose VTE. For instance, patients undergoing radical retropubic prostatectomy have been reported to develop DVT at rates of 1% to 3%, PE at rates of 1% to 3%, and fatal PE at a rate of 0.6%, whereas the incidences of these events are somewhat higher in patients undergoing cystectomy: 8% for DVT, 2% to 4% for PE, and 2% for fatal PE. Radiologic diagnosis of thromboembolism in pelvic surgery patients has yielded higher incidences, with DVT rates of 21% to 51% and PE rates of 11% to 22%.³⁰

Small studies suggest that prophylaxis with either low-dose UFH or LMWH is both effective in reducing VTE risk and safe in urologic cancer surgery patients, although pharmacologic prophylaxis poses a possible increased risk of pelvic hematoma and lymphocele formation in this population.³⁰

Neurosurgery

Most neurosurgical procedures are performed for malignancies. The risk of venography-confirmed VTE in patients undergoing neurosurgery is approximately 30% to 40%.^31,32 Likewise, the risks of intracranial or intraspinal hemorrhage in these patients are high. For this reason, mechanical methods of VTE prophylaxis are preferred in these patients. The use of anticoagulant prophylaxis remains controversial in this setting, although more recent data suggest that it might be safer than previously recognized.

American College of Chest Physicians

National Comprehensive Cancer Network

The National Comprehensive Cancer Network (NCCN) recently published clinical practice guidelines on venous thromboembolic disease in cancer patients.³⁵ The defined at-risk population for these guidelines is the adult cancer inpatient with a diagnosis of (or clinical suspicion for) cancer. The guidelines recommend prophylactic anticoagulation (category 1 recommendation) with or without a sequential compression device as initial prophylaxis, unless the patient has a relative contraindication to anticoagulation, in which case mechanical prophylaxis (sequential compression device or graduated compression stockings) is recommended. (A category 1 recommendation indicates “uniform NCCN consensus, based on high-level evidence.”)

The NCCN guidelines include a specific recommended risk-factor assessment, which includes noting the patient’s age (VTE risk increases beginning at age 40 and then steeply again at age 75), any prior VTE, the presence of familial thrombophilia or active cancer, the use of medications associated with increased VTE risk (chemotherapy, exogenous estrogen compounds, and thalidomide or lenalidomide), and a number of other risk factors for VTE as outlined in the prior two articles in this supplement. The NCCN guidelines explicitly call for assessment of modifiable risk factors for VTE (ie, smoking or other tobacco use, obesity, and a low level of activity or lack of exercise) and call for active patient education on these factors.

American Society of Clinical Oncology

Our recommended algorithm

LINGERING CHALLENGE OF UNDERUTILIZATION

Despite this consensus on ways to reduce thromboembolic risk in this population and the clear evidence of the benefit of VTE prophylaxis in patients with cancer, data from several registries confirm a persistently low utilization of prophylaxis in patients with cancer.^36–38 The global Fundamental Research in Oncology and Thrombosis (FRONTLINE) study surveyed 3,891 clinicians who treat cancer patients regarding their practices with respect to VTE in those patients.36 The survey found that only 52% of respondents routinely used thromboprophylaxis for their surgical patients with cancer. More striking, however, was the finding that most respondents routinely considered thrombo-prophylaxis in only 5% of their medical oncology patients. These data are echoed by findings of other retrospective medical record reviews in patients undergoing major abdominal or abdominothoracic surgery (in many cases for cancer), with VTE prophylaxis rates ranging from 38% to 75%.^37,38

SUMMARY

Patients undergoing surgery for cancer have an increased risk of VTE and fatal PE, even when throm­boprophylaxis is used. Nevertheless, prophylaxis with either LMWH or UFH does reduce venographic VTE event rates in these patients. If UFH is chosen for prophylaxis, a three-times-daily regimen should be used in this population. In specific surgical cancer populations, especially those undergoing abdominal surgery, out-of-hospital prophylaxis with once-daily LMWH is warranted. Current registries reveal that compliance with established guidelines for VTE prophylaxis in this population is low.

Dr. Jaffer: Dr. Amin, based on your study on thrombo-prophylaxis rates in US medical centers, will you comment on rates of prophylaxis for cancer surgery patients?

Dr. Amin: The overall study included approximately 200,000 medical patients and about 80,000 surgical patients enrolled over more than a 3-year period between 2002 and 2005.^39,40 Our goal was to assess rates of prophylaxis and, when it was provided, whether it was appropriate (in terms of type, dosage, and duration) based on the ACCP guidelines. A subanalysis assessed medical cancer patients and surgical cancer patients separately. Medical cancer patients received thromboprophylaxis 56% of the time but received appropriate prophylaxis only 28% of the time. Among surgical cancer patients, appropriate prophylaxis was given only about 24% of the time for those undergoing gynecologic surgery and about 12% of the time for those undergoing neurosurgery. These percentages are consistent with data from other national registries, such as the IMPROVE registry, which documented prophylaxis rates on the order of 45% in medical patients with cancer.⁴¹ We also analyzed the data according to individual practitioners and found that medical oncologists use prophylaxis about 25% of the time, which is relatively consistent with other providers, such as internists and surgeons.

So there is a huge opportunity to improve rates of prophylaxis for this group of patients that national guidelines say are at high risk. Why is prophylaxis so underutilized in the cancer population? One factor may be a misperception about the risk of bleeding with anticoagulants. Yet several studies have shown that the rate of bleeding from prophylaxis is extremely low, whether LMWH or UFH is used, so more awareness of actual bleeding risk is needed. Another factor is the obvious focus among internists and oncologists on treating the patient, with perhaps a reduced consideration of prophylaxis and prevention. A third factor may be a concern about thrombocytopenia. However, in our study of prophylaxis rates in US medical centers, we excluded patients who had thrombocytopenia, yet rates of prophylaxis were still low. Nothing in the literature indicates that anticoagulants cannot be used in patients with platelet counts of 50,000 to 150,000 cells/µL or higher, so this suggests that we need to do more education.

Dr. Jaffer: Dr. Brotman, can you tell us more about how clinicians in practice should use prophylaxis in their neurosurgery patients, such as those undergoing craniotomy or spine surgery for cancer? What is the safest and most efficacious way to prevent DVT in these patients?

Dr. Brotman: First, it’s important to recognize that some sort of prophylaxis needs to be used. Neurosurgery patients are at an extremely high risk for thromboembolic events, and such events are often fatal in these patients. Having said that, the jury is still out on whether the prophylaxis in these patients should be compression devices or anticoagulation. This gives physicians some latitude in their decisions. They can decide not to use pharmacologic prophylaxis so long as they use pneumatic compression devices consistently, perhaps even starting during the operation and certainly throughout hospitalization when the patient is immobilized.

Certainly, the concerns about using full-dose anticoagulation in the immediate postoperative setting in neurosurgery patients are valid. Yet these patients are at very high risk for thromboembolic events, and if we take too cautious an approach to prophylaxis in the immediate perioperative setting, more patients are going to have thromboembolic events, at which point management decisions become much more difficult. The risk of intracranial bleeding with anticoagulation to treat a patient who develops a DVT at postoperative day 10 will certainly be higher than it would have been with lower-dose perioperative prophylactic anticoagulation. Plus, if you put in a filter at that point, the outcomes tend to be poor. Therefore, I believe there is some degree of risk that we should be willing to take with regard to perioperative bleeding, even in neurosurgery patients.

Dr. McKean: I’d like to make a point about combination prophylaxis. At many institutions, compression stockings and sequential compression devices are used preoperatively and intraoperatively, and then pharmacologic prophylaxis—for example, twice-daily UFH—is used postoperatively. There is concern that these patients are hypercoagulable, and most clinicians believe that mechanical prophylaxis alone, even with sequential compression devices plus compression stockings, is not aggressive enough in these high-risk patients.

Dr. Jaffer: Dr. Spyropoulos, what is the optimal duration of pharmacologic prophylaxis for cancer surgery patients?

Dr. Spyropoulos: First let’s consider in-hospital prophylaxis. The supportive data for in-hospital prophylaxis are strong, and the duration of therapy used in the major in-hospital prophylaxis trials was 7 to 10 days. With regard to extended prophylaxis, we have at least two moderately sized randomized controlled trials, ENOXACAN II²³ and the substudy of FAME,²⁴ that demonstrated that extending prophylaxis with LMWH at doses of 3,400 U once daily (5,000 IU of dalteparin; 40 mg of enoxaparin) reduces VTE risk at postoperative day 30. Also, recent data from the @RISTOS registry show that in cancer surgery patients, especially those having abdominal or pelvic procedures, the leading cause of 30-day mortality was VTE.⁸ This registry also shows that despite prophylaxis, the rate of symptomatic VTE can be as high as 2%, with the rate of fatal VTE approaching 1%. Thus, in cancer patients undergoing abdominal or pelvic surgery, physicians should strongly consider prophylaxis of up to 30 days’ duration.

Dr. Jaffer: One striking finding from the @RISTOS registry was that 40% of VTE events in these cancer surgery patients occurred after postoperative day 21. This really underscores the need to consider prophylaxis for at least 4 weeks in these patients in real-world practice.

Dr. Brotman: The other striking finding from that registry was that the in-hospital prophylaxis rate was quite high, about 80%, and the rate of extended prophylaxis approached 35%. These are rates that are rarely achieved in clinical practice. Yet despite these high levels of prophylaxis, patients in this registry still had a high incidence of morbidity and mortality from VTE. This suggests that we need to improve our out-of-hospital VTE prevention paradigms.

Dr. Jaffer: Dr. Deitelzweig, oncologists and internists are often unsure about whether their ambulatory cancer patients who are receiving chemotherapy should be on any form of prophylaxis. What is your opinion?

Dr. Deitelzweig: That question comes up regularly because these patients are encountered across many medical specialties. At this point, all of the large organizations, including ASCO and NCCN, are advocating that prophylaxis is not indicated for such patients.

Venous thromboembolism (VTE) is a major complication of cancer, occurring in 4% to 20% of patients,¹ and is one of the leading causes of death in cancer patients, although these figures are believed to be underestimates, given the low autopsy rates among cancer patients.² In hospitalized cancer patients specifically, VTE is the second leading cause of death.^3,4 The risk of VTE in cancer patients undergoing surgery is three to five times greater than that in surgical patients without cancer.⁴ Moreover, cancer patients with symptomatic deep vein thrombosis (DVT) exhibit a high risk of recurrent VTE that may persist for many years after the index event.⁵

VTE PREVENTION POSES PARTICULAR CHALLENGES IN CANCER PATIENTS

Until recently, data on VTE prevention specific to cancer patients have been sparse. Cancer patients have represented only a small subset (< 20%) of participants in most of the largest clinical trials of VTE prophylaxis. Until the past 2 or 3 years, clinicians largely have had to extrapolate their approach to VTE prophylaxis in cancer patients from data in patients without cancer, bearing in mind that cancer patients are among the populations at highest risk of developing VTE.

High rates of VTE, even with prophylaxis

Further insights have come from the @RISTOS project, a Web-based prospective registry of patients undergoing general, urologic, or gynecologic surgery for cancer at multiple centers in Italy.⁸ Of the 2,372 patients tracked in this study, 82% received in-hospital VTE prophylaxis and 31% received prophylaxis following discharge. Despite this relatively high frequency of prophylaxis, however, the incidence of clinically overt VTE was 2.1% and the incidence of fatal VTE was 0.8%. Notably, most VTE events occurred after hospital discharge, and VTE was the most common cause of 30-day postoperative death in this registry.

RISK FACTORS: CANCER TYPE AND TREATMENT LOOM LARGE

Both the type and stage of a patient’s cancer are important in assessing the risk of VTE. For men, cancers of the prostate, colon, brain, and lung have been associated with an increased risk of VTE. Among women, cancers of the breast, ovary, and lung have been especially implicated as risk factors for VTE.^9,10

The type of cancer therapy also influences VTE risk:

• Surgery. Among patients who undergo cancer-related surgery, the rate of proximal DVT is 10% to 20%, the rate of clinically evident PE is 4% to 10%, and the incidence of fatal PE is 0.2% to 5%.^8,11
• Systemic treatments, including chemotherapy and hormone therapy, are also associated with an increased risk of VTE.^12–15
• Central venous catheters. Approximately 4% of cancer patients who have central venous catheters placed develop clinically relevant VTE.^16,17

In addition to the above risks related to cancer treatments, the following have been identified as risk factors for VTE in surgical oncology patients:

• Age greater than 40 years (risk also increases steeply after age 60 and again after age 75)
• Cancer procoagulants
• Thrombophilia
• Length and complications of cancer surgery (ie, often involving tissue trauma and immobilization)
• Debilitation and slow recovery.

Another risk factor worth noting is perioperative transfusion, as illustrated in a recent study of 14,104 adults undergoing colorectal cancer resection.¹⁸ The overall incidence of VTE in these patients was 1.0%, and the risk of death was nearly four times as great in patients who developed VTE as in those who did not. Notably, the need for transfusion was a marker of increased risk of VTE, particularly in women: women who received perioperative transfusions had almost double the risk of developing VTE compared with women who did not receive transfusions (P = .004).

CLINICAL TRIALS OF PROPHYLAXIS IN CANCER SURGERY PATIENTS

LMWH vs UFH for in-hospital prophylaxis

Two large randomized, double-blind trials have compared low-molecular-weight heparin (LMWH) with low-dose unfractionated heparin (UFH) for VTE prophylaxis in surgical patients with cancer—the Enoxaparin and Cancer (ENOXACAN) study¹⁹ and the Canadian Colorectal Surgery DVT Prophylaxis Trial.²⁰ Patients in these studies underwent surgery for abdominal or pelvic cancer (mostly colorectal cancer). Both studies compared 40 mg of the LMWH enoxaparin given once daily with 5,000 U of UFH given three times daily for 7 to 10 days postoperatively. Outcome measures were the presence of DVT determined by venography on day 7 to 10 and the incidence of symptomatic VTE. Rates of VTE were statistically equivalent between the two treatment arms in both ENOXACAN (14.7% with LMWH vs 18.2% with UFH) and the Canadian Colorectal Surgery study (9.4% with both therapies), as were rates of major bleeding (4.1% with LMWH vs 2.9% with UFH in ENOXACAN; 2.7% with LMWH vs 1.5% with UFH in the Canadian study).

These findings are consistent with a 2001 meta-analysis by Mismetti et al of all available randomized trials comparing LMWH with placebo or with UFH for VTE prophylaxis in general surgery.²¹ This analysis found no differences in rates of asymptomatic DVT, clinical PE, clinical thromboembolism, death, major hemorrhage, total hemorrhage, wound hematoma, or need for transfusion between LMWH and UFH in patients undergoing either cancer-related surgery or surgery not related to cancer.

Fondaparinux for in-hospital prophylaxis

Subgroup analysis of the large randomized trial known as PEGASUS²² sheds some light on the efficacy of the factor Xa inhibitor fondaparinux relative to LMWH for thromboprophylaxis in cancer surgery patients. PEGASUS compared fondaparinux 2.5 mg once daily with the LMWH dalteparin 5,000 IU once daily for 5 to 9 days in patients undergoing high-risk abdominal surgery. Among the study’s 1,408 patients undergoing surgery for cancer, rates of VTE were 4.7% in the fondaparinux group compared with 7.7% in the LMWH group, a relative risk reduction of 38.6% with fondaparinux (95% CI, 6.7% to 59.6%). In contrast, in the rest of the PEGASUS population (patients undergoing abdominal surgery for reasons other than cancer), LMWH was nonsignificantly more efficacious at preventing VTE than was fondaparinux. Rates of major bleeding in this cancer subgroup were comparable between the two treatments.

Two additional randomized trials have evaluated extended prophylaxis with LMWH in surgical cancer patients—ENOXACAN II²³ and the Fragmin After Major Abdominal Surgery (FAME) study.²⁴

In ENOXACAN II, patients undergoing surgery for abdominal or pelvic cancer first received 6 to 10 days of prophylaxis with enoxaparin 40 mg once daily and then were randomized in a double-blind fashion to an additional 21 days of enoxaparin or placebo.²³ Among 332 patients in the intent-to-treat analysis, the rate of VTE at the end of the double-blind phase was reduced from 12.0% with placebo to 4.8% with extended-duration enoxaparin (P = .02), an effect that was maintained at 3-month follow-up (P = .01). There was no significant difference between the two groups in rates of major bleeding events or any bleeding events.

In FAME, patients received 5,000 IU of dalteparin once daily for 1 week following major abdominal surgery and then were randomized in open-label fashion to either placebo or extended prophylaxis with dalteparin for 3 more weeks; a subanalysis examined outcomes in the 198 FAME participants whose abdominal surgery was for cancer.²⁴ Among these 198 cancer surgery patients, the rate of venography-documented VTE at 4 weeks was reduced from 19.6% with placebo to 8.8% with extended-duration dalteparin, a relative reduction of 55% (P = .03). The rate of proximal DVT was reduced from 10.4% to 2.2% with extended prophylaxis, a relative reduction of 79% (P = .02).

The number needed to treat with extended LMWH prophylaxis to prevent one VTE event was 14 in ENOXACAN II²³ and 9 in the FAME subanalysis of cancer surgery patients.²⁴

New systematic review of relevant trials

Leonardi et al recently published a systematic review of 26 randomized controlled trials of DVT prophylaxis in 7,639 cancer surgery patients.²⁵ They found the overall incidence of DVT to be 12.7% in those who received pharmacologic prophylaxis compared with 35.2% in controls. They also found high-dose LMWH therapy (> 3,400 U daily) to be associated with a significantly lower incidence of DVT than low-dose LMWH therapy (≤ 3,400 U daily) (7.9% vs 14.5%, respectively; P < .01). No differences were demonstrated between LMWH and UFH in preventing DVT, DVT location, or bleeding. Bleeding complications requiring discontinuation of pharmacologic prophylaxis occurred in 3% of patients overall.

Implications of HIT

The sequelae of heparin-induced thrombocytopenia (HIT) can have major consequences for cancer surgery patients. The incidence of HIT is markedly lower with LMWH than with UFH, as demonstrated in a nested case-control study by Creekmore et al.²⁶ These researchers also found that the average cost of an admission during which HIT developed was nearly four times as great as the average cost of an admission during which UFH or LMWH was given without development of HIT ($56,364 vs $15,231; P < .001).

EVIDENCE IN SPECIFIC ONCOLOGIC POPULATIONS

Most of the patients in the trials reviewed above underwent abdominal surgery for malignancy. Although studies of VTE prophylaxis in patients undergoing nonabdominal cancer surgery are relatively few, some data are available for a few other specific oncologic populations, as reviewed below.

Surgery for gynecologic cancer

There is a paucity of randomized controlled trials or prospective observational studies on VTE and its prevention in the gynecologic cancer surgery population. Based on small historical studies, the postoperative risk of VTE in this population varies from 12% to 35%.^27,28 Twice-daily administration of UFH 5,000 U appears to be ineffective as VTE prophylaxis in this population, but increasing the frequency to three times daily reduces VTE risk by 50% to 60% compared with placebo. Once-daily LMWH is comparable to three-times-daily UFH in efficacy and safety in this population.

A systematic Cochrane review of eight randomized controlled trials in patients undergoing major gynecologic surgery revealed that heparin prophylaxis (either UFH or LMWH) reduces the risk of DVT by 70% compared with no prophylaxis, with an identical risk reduction specifically among women with malignancy (odds ratio, 0.30; 95% CI, 0.10 to 0.89).²⁹ This review found no evidence that anticoagulation reduces the risk of PE following major gynecologic surgery. LMWH and UFH were similar in efficacy for preventing DVT and had a comparable risk of bleeding complications.

Surgery for urologic cancer

The risk of VTE and the benefits of thromboprophylaxis also are poorly studied in patients undergoing surgery for urologic cancer.

The risk of VTE varies with the type of urologic surgery and the method used to diagnose VTE. For instance, patients undergoing radical retropubic prostatectomy have been reported to develop DVT at rates of 1% to 3%, PE at rates of 1% to 3%, and fatal PE at a rate of 0.6%, whereas the incidences of these events are somewhat higher in patients undergoing cystectomy: 8% for DVT, 2% to 4% for PE, and 2% for fatal PE. Radiologic diagnosis of thromboembolism in pelvic surgery patients has yielded higher incidences, with DVT rates of 21% to 51% and PE rates of 11% to 22%.³⁰

Small studies suggest that prophylaxis with either low-dose UFH or LMWH is both effective in reducing VTE risk and safe in urologic cancer surgery patients, although pharmacologic prophylaxis poses a possible increased risk of pelvic hematoma and lymphocele formation in this population.³⁰

Neurosurgery

Most neurosurgical procedures are performed for malignancies. The risk of venography-confirmed VTE in patients undergoing neurosurgery is approximately 30% to 40%.^31,32 Likewise, the risks of intracranial or intraspinal hemorrhage in these patients are high. For this reason, mechanical methods of VTE prophylaxis are preferred in these patients. The use of anticoagulant prophylaxis remains controversial in this setting, although more recent data suggest that it might be safer than previously recognized.

American College of Chest Physicians

National Comprehensive Cancer Network

The National Comprehensive Cancer Network (NCCN) recently published clinical practice guidelines on venous thromboembolic disease in cancer patients.³⁵ The defined at-risk population for these guidelines is the adult cancer inpatient with a diagnosis of (or clinical suspicion for) cancer. The guidelines recommend prophylactic anticoagulation (category 1 recommendation) with or without a sequential compression device as initial prophylaxis, unless the patient has a relative contraindication to anticoagulation, in which case mechanical prophylaxis (sequential compression device or graduated compression stockings) is recommended. (A category 1 recommendation indicates “uniform NCCN consensus, based on high-level evidence.”)

The NCCN guidelines include a specific recommended risk-factor assessment, which includes noting the patient’s age (VTE risk increases beginning at age 40 and then steeply again at age 75), any prior VTE, the presence of familial thrombophilia or active cancer, the use of medications associated with increased VTE risk (chemotherapy, exogenous estrogen compounds, and thalidomide or lenalidomide), and a number of other risk factors for VTE as outlined in the prior two articles in this supplement. The NCCN guidelines explicitly call for assessment of modifiable risk factors for VTE (ie, smoking or other tobacco use, obesity, and a low level of activity or lack of exercise) and call for active patient education on these factors.

American Society of Clinical Oncology

Our recommended algorithm

LINGERING CHALLENGE OF UNDERUTILIZATION

Despite this consensus on ways to reduce thromboembolic risk in this population and the clear evidence of the benefit of VTE prophylaxis in patients with cancer, data from several registries confirm a persistently low utilization of prophylaxis in patients with cancer.^36–38 The global Fundamental Research in Oncology and Thrombosis (FRONTLINE) study surveyed 3,891 clinicians who treat cancer patients regarding their practices with respect to VTE in those patients.36 The survey found that only 52% of respondents routinely used thromboprophylaxis for their surgical patients with cancer. More striking, however, was the finding that most respondents routinely considered thrombo-prophylaxis in only 5% of their medical oncology patients. These data are echoed by findings of other retrospective medical record reviews in patients undergoing major abdominal or abdominothoracic surgery (in many cases for cancer), with VTE prophylaxis rates ranging from 38% to 75%.^37,38

SUMMARY

Patients undergoing surgery for cancer have an increased risk of VTE and fatal PE, even when throm­boprophylaxis is used. Nevertheless, prophylaxis with either LMWH or UFH does reduce venographic VTE event rates in these patients. If UFH is chosen for prophylaxis, a three-times-daily regimen should be used in this population. In specific surgical cancer populations, especially those undergoing abdominal surgery, out-of-hospital prophylaxis with once-daily LMWH is warranted. Current registries reveal that compliance with established guidelines for VTE prophylaxis in this population is low.

Dr. Jaffer: Dr. Amin, based on your study on thrombo-prophylaxis rates in US medical centers, will you comment on rates of prophylaxis for cancer surgery patients?

Dr. Amin: The overall study included approximately 200,000 medical patients and about 80,000 surgical patients enrolled over more than a 3-year period between 2002 and 2005.^39,40 Our goal was to assess rates of prophylaxis and, when it was provided, whether it was appropriate (in terms of type, dosage, and duration) based on the ACCP guidelines. A subanalysis assessed medical cancer patients and surgical cancer patients separately. Medical cancer patients received thromboprophylaxis 56% of the time but received appropriate prophylaxis only 28% of the time. Among surgical cancer patients, appropriate prophylaxis was given only about 24% of the time for those undergoing gynecologic surgery and about 12% of the time for those undergoing neurosurgery. These percentages are consistent with data from other national registries, such as the IMPROVE registry, which documented prophylaxis rates on the order of 45% in medical patients with cancer.⁴¹ We also analyzed the data according to individual practitioners and found that medical oncologists use prophylaxis about 25% of the time, which is relatively consistent with other providers, such as internists and surgeons.

So there is a huge opportunity to improve rates of prophylaxis for this group of patients that national guidelines say are at high risk. Why is prophylaxis so underutilized in the cancer population? One factor may be a misperception about the risk of bleeding with anticoagulants. Yet several studies have shown that the rate of bleeding from prophylaxis is extremely low, whether LMWH or UFH is used, so more awareness of actual bleeding risk is needed. Another factor is the obvious focus among internists and oncologists on treating the patient, with perhaps a reduced consideration of prophylaxis and prevention. A third factor may be a concern about thrombocytopenia. However, in our study of prophylaxis rates in US medical centers, we excluded patients who had thrombocytopenia, yet rates of prophylaxis were still low. Nothing in the literature indicates that anticoagulants cannot be used in patients with platelet counts of 50,000 to 150,000 cells/µL or higher, so this suggests that we need to do more education.

Dr. Jaffer: Dr. Brotman, can you tell us more about how clinicians in practice should use prophylaxis in their neurosurgery patients, such as those undergoing craniotomy or spine surgery for cancer? What is the safest and most efficacious way to prevent DVT in these patients?

Dr. Brotman: First, it’s important to recognize that some sort of prophylaxis needs to be used. Neurosurgery patients are at an extremely high risk for thromboembolic events, and such events are often fatal in these patients. Having said that, the jury is still out on whether the prophylaxis in these patients should be compression devices or anticoagulation. This gives physicians some latitude in their decisions. They can decide not to use pharmacologic prophylaxis so long as they use pneumatic compression devices consistently, perhaps even starting during the operation and certainly throughout hospitalization when the patient is immobilized.

Certainly, the concerns about using full-dose anticoagulation in the immediate postoperative setting in neurosurgery patients are valid. Yet these patients are at very high risk for thromboembolic events, and if we take too cautious an approach to prophylaxis in the immediate perioperative setting, more patients are going to have thromboembolic events, at which point management decisions become much more difficult. The risk of intracranial bleeding with anticoagulation to treat a patient who develops a DVT at postoperative day 10 will certainly be higher than it would have been with lower-dose perioperative prophylactic anticoagulation. Plus, if you put in a filter at that point, the outcomes tend to be poor. Therefore, I believe there is some degree of risk that we should be willing to take with regard to perioperative bleeding, even in neurosurgery patients.

Dr. McKean: I’d like to make a point about combination prophylaxis. At many institutions, compression stockings and sequential compression devices are used preoperatively and intraoperatively, and then pharmacologic prophylaxis—for example, twice-daily UFH—is used postoperatively. There is concern that these patients are hypercoagulable, and most clinicians believe that mechanical prophylaxis alone, even with sequential compression devices plus compression stockings, is not aggressive enough in these high-risk patients.

Dr. Jaffer: Dr. Spyropoulos, what is the optimal duration of pharmacologic prophylaxis for cancer surgery patients?

Dr. Spyropoulos: First let’s consider in-hospital prophylaxis. The supportive data for in-hospital prophylaxis are strong, and the duration of therapy used in the major in-hospital prophylaxis trials was 7 to 10 days. With regard to extended prophylaxis, we have at least two moderately sized randomized controlled trials, ENOXACAN II²³ and the substudy of FAME,²⁴ that demonstrated that extending prophylaxis with LMWH at doses of 3,400 U once daily (5,000 IU of dalteparin; 40 mg of enoxaparin) reduces VTE risk at postoperative day 30. Also, recent data from the @RISTOS registry show that in cancer surgery patients, especially those having abdominal or pelvic procedures, the leading cause of 30-day mortality was VTE.⁸ This registry also shows that despite prophylaxis, the rate of symptomatic VTE can be as high as 2%, with the rate of fatal VTE approaching 1%. Thus, in cancer patients undergoing abdominal or pelvic surgery, physicians should strongly consider prophylaxis of up to 30 days’ duration.

Dr. Jaffer: One striking finding from the @RISTOS registry was that 40% of VTE events in these cancer surgery patients occurred after postoperative day 21. This really underscores the need to consider prophylaxis for at least 4 weeks in these patients in real-world practice.

Dr. Brotman: The other striking finding from that registry was that the in-hospital prophylaxis rate was quite high, about 80%, and the rate of extended prophylaxis approached 35%. These are rates that are rarely achieved in clinical practice. Yet despite these high levels of prophylaxis, patients in this registry still had a high incidence of morbidity and mortality from VTE. This suggests that we need to improve our out-of-hospital VTE prevention paradigms.

Dr. Jaffer: Dr. Deitelzweig, oncologists and internists are often unsure about whether their ambulatory cancer patients who are receiving chemotherapy should be on any form of prophylaxis. What is your opinion?

Dr. Deitelzweig: That question comes up regularly because these patients are encountered across many medical specialties. At this point, all of the large organizations, including ASCO and NCCN, are advocating that prophylaxis is not indicated for such patients.

Venous thromboembolism (VTE) is a major complication of cancer, occurring in 4% to 20% of patients,¹ and is one of the leading causes of death in cancer patients, although these figures are believed to be underestimates, given the low autopsy rates among cancer patients.² In hospitalized cancer patients specifically, VTE is the second leading cause of death.^3,4 The risk of VTE in cancer patients undergoing surgery is three to five times greater than that in surgical patients without cancer.⁴ Moreover, cancer patients with symptomatic deep vein thrombosis (DVT) exhibit a high risk of recurrent VTE that may persist for many years after the index event.⁵

VTE PREVENTION POSES PARTICULAR CHALLENGES IN CANCER PATIENTS

Until recently, data on VTE prevention specific to cancer patients have been sparse. Cancer patients have represented only a small subset (< 20%) of participants in most of the largest clinical trials of VTE prophylaxis. Until the past 2 or 3 years, clinicians largely have had to extrapolate their approach to VTE prophylaxis in cancer patients from data in patients without cancer, bearing in mind that cancer patients are among the populations at highest risk of developing VTE.

High rates of VTE, even with prophylaxis

Further insights have come from the @RISTOS project, a Web-based prospective registry of patients undergoing general, urologic, or gynecologic surgery for cancer at multiple centers in Italy.⁸ Of the 2,372 patients tracked in this study, 82% received in-hospital VTE prophylaxis and 31% received prophylaxis following discharge. Despite this relatively high frequency of prophylaxis, however, the incidence of clinically overt VTE was 2.1% and the incidence of fatal VTE was 0.8%. Notably, most VTE events occurred after hospital discharge, and VTE was the most common cause of 30-day postoperative death in this registry.

RISK FACTORS: CANCER TYPE AND TREATMENT LOOM LARGE

Both the type and stage of a patient’s cancer are important in assessing the risk of VTE. For men, cancers of the prostate, colon, brain, and lung have been associated with an increased risk of VTE. Among women, cancers of the breast, ovary, and lung have been especially implicated as risk factors for VTE.^9,10

The type of cancer therapy also influences VTE risk:

• Surgery. Among patients who undergo cancer-related surgery, the rate of proximal DVT is 10% to 20%, the rate of clinically evident PE is 4% to 10%, and the incidence of fatal PE is 0.2% to 5%.^8,11
• Systemic treatments, including chemotherapy and hormone therapy, are also associated with an increased risk of VTE.^12–15
• Central venous catheters. Approximately 4% of cancer patients who have central venous catheters placed develop clinically relevant VTE.^16,17

In addition to the above risks related to cancer treatments, the following have been identified as risk factors for VTE in surgical oncology patients:

• Age greater than 40 years (risk also increases steeply after age 60 and again after age 75)
• Cancer procoagulants
• Thrombophilia
• Length and complications of cancer surgery (ie, often involving tissue trauma and immobilization)
• Debilitation and slow recovery.

Another risk factor worth noting is perioperative transfusion, as illustrated in a recent study of 14,104 adults undergoing colorectal cancer resection.¹⁸ The overall incidence of VTE in these patients was 1.0%, and the risk of death was nearly four times as great in patients who developed VTE as in those who did not. Notably, the need for transfusion was a marker of increased risk of VTE, particularly in women: women who received perioperative transfusions had almost double the risk of developing VTE compared with women who did not receive transfusions (P = .004).

CLINICAL TRIALS OF PROPHYLAXIS IN CANCER SURGERY PATIENTS

LMWH vs UFH for in-hospital prophylaxis

Two large randomized, double-blind trials have compared low-molecular-weight heparin (LMWH) with low-dose unfractionated heparin (UFH) for VTE prophylaxis in surgical patients with cancer—the Enoxaparin and Cancer (ENOXACAN) study¹⁹ and the Canadian Colorectal Surgery DVT Prophylaxis Trial.²⁰ Patients in these studies underwent surgery for abdominal or pelvic cancer (mostly colorectal cancer). Both studies compared 40 mg of the LMWH enoxaparin given once daily with 5,000 U of UFH given three times daily for 7 to 10 days postoperatively. Outcome measures were the presence of DVT determined by venography on day 7 to 10 and the incidence of symptomatic VTE. Rates of VTE were statistically equivalent between the two treatment arms in both ENOXACAN (14.7% with LMWH vs 18.2% with UFH) and the Canadian Colorectal Surgery study (9.4% with both therapies), as were rates of major bleeding (4.1% with LMWH vs 2.9% with UFH in ENOXACAN; 2.7% with LMWH vs 1.5% with UFH in the Canadian study).

These findings are consistent with a 2001 meta-analysis by Mismetti et al of all available randomized trials comparing LMWH with placebo or with UFH for VTE prophylaxis in general surgery.²¹ This analysis found no differences in rates of asymptomatic DVT, clinical PE, clinical thromboembolism, death, major hemorrhage, total hemorrhage, wound hematoma, or need for transfusion between LMWH and UFH in patients undergoing either cancer-related surgery or surgery not related to cancer.

Fondaparinux for in-hospital prophylaxis

Subgroup analysis of the large randomized trial known as PEGASUS²² sheds some light on the efficacy of the factor Xa inhibitor fondaparinux relative to LMWH for thromboprophylaxis in cancer surgery patients. PEGASUS compared fondaparinux 2.5 mg once daily with the LMWH dalteparin 5,000 IU once daily for 5 to 9 days in patients undergoing high-risk abdominal surgery. Among the study’s 1,408 patients undergoing surgery for cancer, rates of VTE were 4.7% in the fondaparinux group compared with 7.7% in the LMWH group, a relative risk reduction of 38.6% with fondaparinux (95% CI, 6.7% to 59.6%). In contrast, in the rest of the PEGASUS population (patients undergoing abdominal surgery for reasons other than cancer), LMWH was nonsignificantly more efficacious at preventing VTE than was fondaparinux. Rates of major bleeding in this cancer subgroup were comparable between the two treatments.

Two additional randomized trials have evaluated extended prophylaxis with LMWH in surgical cancer patients—ENOXACAN II²³ and the Fragmin After Major Abdominal Surgery (FAME) study.²⁴

In ENOXACAN II, patients undergoing surgery for abdominal or pelvic cancer first received 6 to 10 days of prophylaxis with enoxaparin 40 mg once daily and then were randomized in a double-blind fashion to an additional 21 days of enoxaparin or placebo.²³ Among 332 patients in the intent-to-treat analysis, the rate of VTE at the end of the double-blind phase was reduced from 12.0% with placebo to 4.8% with extended-duration enoxaparin (P = .02), an effect that was maintained at 3-month follow-up (P = .01). There was no significant difference between the two groups in rates of major bleeding events or any bleeding events.

In FAME, patients received 5,000 IU of dalteparin once daily for 1 week following major abdominal surgery and then were randomized in open-label fashion to either placebo or extended prophylaxis with dalteparin for 3 more weeks; a subanalysis examined outcomes in the 198 FAME participants whose abdominal surgery was for cancer.²⁴ Among these 198 cancer surgery patients, the rate of venography-documented VTE at 4 weeks was reduced from 19.6% with placebo to 8.8% with extended-duration dalteparin, a relative reduction of 55% (P = .03). The rate of proximal DVT was reduced from 10.4% to 2.2% with extended prophylaxis, a relative reduction of 79% (P = .02).

The number needed to treat with extended LMWH prophylaxis to prevent one VTE event was 14 in ENOXACAN II²³ and 9 in the FAME subanalysis of cancer surgery patients.²⁴

New systematic review of relevant trials

Leonardi et al recently published a systematic review of 26 randomized controlled trials of DVT prophylaxis in 7,639 cancer surgery patients.²⁵ They found the overall incidence of DVT to be 12.7% in those who received pharmacologic prophylaxis compared with 35.2% in controls. They also found high-dose LMWH therapy (> 3,400 U daily) to be associated with a significantly lower incidence of DVT than low-dose LMWH therapy (≤ 3,400 U daily) (7.9% vs 14.5%, respectively; P < .01). No differences were demonstrated between LMWH and UFH in preventing DVT, DVT location, or bleeding. Bleeding complications requiring discontinuation of pharmacologic prophylaxis occurred in 3% of patients overall.

Implications of HIT

The sequelae of heparin-induced thrombocytopenia (HIT) can have major consequences for cancer surgery patients. The incidence of HIT is markedly lower with LMWH than with UFH, as demonstrated in a nested case-control study by Creekmore et al.²⁶ These researchers also found that the average cost of an admission during which HIT developed was nearly four times as great as the average cost of an admission during which UFH or LMWH was given without development of HIT ($56,364 vs $15,231; P < .001).

EVIDENCE IN SPECIFIC ONCOLOGIC POPULATIONS

Most of the patients in the trials reviewed above underwent abdominal surgery for malignancy. Although studies of VTE prophylaxis in patients undergoing nonabdominal cancer surgery are relatively few, some data are available for a few other specific oncologic populations, as reviewed below.

Surgery for gynecologic cancer

There is a paucity of randomized controlled trials or prospective observational studies on VTE and its prevention in the gynecologic cancer surgery population. Based on small historical studies, the postoperative risk of VTE in this population varies from 12% to 35%.^27,28 Twice-daily administration of UFH 5,000 U appears to be ineffective as VTE prophylaxis in this population, but increasing the frequency to three times daily reduces VTE risk by 50% to 60% compared with placebo. Once-daily LMWH is comparable to three-times-daily UFH in efficacy and safety in this population.

A systematic Cochrane review of eight randomized controlled trials in patients undergoing major gynecologic surgery revealed that heparin prophylaxis (either UFH or LMWH) reduces the risk of DVT by 70% compared with no prophylaxis, with an identical risk reduction specifically among women with malignancy (odds ratio, 0.30; 95% CI, 0.10 to 0.89).²⁹ This review found no evidence that anticoagulation reduces the risk of PE following major gynecologic surgery. LMWH and UFH were similar in efficacy for preventing DVT and had a comparable risk of bleeding complications.

Surgery for urologic cancer

The risk of VTE and the benefits of thromboprophylaxis also are poorly studied in patients undergoing surgery for urologic cancer.

The risk of VTE varies with the type of urologic surgery and the method used to diagnose VTE. For instance, patients undergoing radical retropubic prostatectomy have been reported to develop DVT at rates of 1% to 3%, PE at rates of 1% to 3%, and fatal PE at a rate of 0.6%, whereas the incidences of these events are somewhat higher in patients undergoing cystectomy: 8% for DVT, 2% to 4% for PE, and 2% for fatal PE. Radiologic diagnosis of thromboembolism in pelvic surgery patients has yielded higher incidences, with DVT rates of 21% to 51% and PE rates of 11% to 22%.³⁰

Small studies suggest that prophylaxis with either low-dose UFH or LMWH is both effective in reducing VTE risk and safe in urologic cancer surgery patients, although pharmacologic prophylaxis poses a possible increased risk of pelvic hematoma and lymphocele formation in this population.³⁰

Neurosurgery

Most neurosurgical procedures are performed for malignancies. The risk of venography-confirmed VTE in patients undergoing neurosurgery is approximately 30% to 40%.^31,32 Likewise, the risks of intracranial or intraspinal hemorrhage in these patients are high. For this reason, mechanical methods of VTE prophylaxis are preferred in these patients. The use of anticoagulant prophylaxis remains controversial in this setting, although more recent data suggest that it might be safer than previously recognized.

American College of Chest Physicians

National Comprehensive Cancer Network

The National Comprehensive Cancer Network (NCCN) recently published clinical practice guidelines on venous thromboembolic disease in cancer patients.³⁵ The defined at-risk population for these guidelines is the adult cancer inpatient with a diagnosis of (or clinical suspicion for) cancer. The guidelines recommend prophylactic anticoagulation (category 1 recommendation) with or without a sequential compression device as initial prophylaxis, unless the patient has a relative contraindication to anticoagulation, in which case mechanical prophylaxis (sequential compression device or graduated compression stockings) is recommended. (A category 1 recommendation indicates “uniform NCCN consensus, based on high-level evidence.”)

The NCCN guidelines include a specific recommended risk-factor assessment, which includes noting the patient’s age (VTE risk increases beginning at age 40 and then steeply again at age 75), any prior VTE, the presence of familial thrombophilia or active cancer, the use of medications associated with increased VTE risk (chemotherapy, exogenous estrogen compounds, and thalidomide or lenalidomide), and a number of other risk factors for VTE as outlined in the prior two articles in this supplement. The NCCN guidelines explicitly call for assessment of modifiable risk factors for VTE (ie, smoking or other tobacco use, obesity, and a low level of activity or lack of exercise) and call for active patient education on these factors.

American Society of Clinical Oncology

Our recommended algorithm

LINGERING CHALLENGE OF UNDERUTILIZATION

Despite this consensus on ways to reduce thromboembolic risk in this population and the clear evidence of the benefit of VTE prophylaxis in patients with cancer, data from several registries confirm a persistently low utilization of prophylaxis in patients with cancer.^36–38 The global Fundamental Research in Oncology and Thrombosis (FRONTLINE) study surveyed 3,891 clinicians who treat cancer patients regarding their practices with respect to VTE in those patients.36 The survey found that only 52% of respondents routinely used thromboprophylaxis for their surgical patients with cancer. More striking, however, was the finding that most respondents routinely considered thrombo-prophylaxis in only 5% of their medical oncology patients. These data are echoed by findings of other retrospective medical record reviews in patients undergoing major abdominal or abdominothoracic surgery (in many cases for cancer), with VTE prophylaxis rates ranging from 38% to 75%.^37,38

SUMMARY

Patients undergoing surgery for cancer have an increased risk of VTE and fatal PE, even when throm­boprophylaxis is used. Nevertheless, prophylaxis with either LMWH or UFH does reduce venographic VTE event rates in these patients. If UFH is chosen for prophylaxis, a three-times-daily regimen should be used in this population. In specific surgical cancer populations, especially those undergoing abdominal surgery, out-of-hospital prophylaxis with once-daily LMWH is warranted. Current registries reveal that compliance with established guidelines for VTE prophylaxis in this population is low.

Dr. Jaffer: Dr. Amin, based on your study on thrombo-prophylaxis rates in US medical centers, will you comment on rates of prophylaxis for cancer surgery patients?

Dr. Amin: The overall study included approximately 200,000 medical patients and about 80,000 surgical patients enrolled over more than a 3-year period between 2002 and 2005.^39,40 Our goal was to assess rates of prophylaxis and, when it was provided, whether it was appropriate (in terms of type, dosage, and duration) based on the ACCP guidelines. A subanalysis assessed medical cancer patients and surgical cancer patients separately. Medical cancer patients received thromboprophylaxis 56% of the time but received appropriate prophylaxis only 28% of the time. Among surgical cancer patients, appropriate prophylaxis was given only about 24% of the time for those undergoing gynecologic surgery and about 12% of the time for those undergoing neurosurgery. These percentages are consistent with data from other national registries, such as the IMPROVE registry, which documented prophylaxis rates on the order of 45% in medical patients with cancer.⁴¹ We also analyzed the data according to individual practitioners and found that medical oncologists use prophylaxis about 25% of the time, which is relatively consistent with other providers, such as internists and surgeons.

So there is a huge opportunity to improve rates of prophylaxis for this group of patients that national guidelines say are at high risk. Why is prophylaxis so underutilized in the cancer population? One factor may be a misperception about the risk of bleeding with anticoagulants. Yet several studies have shown that the rate of bleeding from prophylaxis is extremely low, whether LMWH or UFH is used, so more awareness of actual bleeding risk is needed. Another factor is the obvious focus among internists and oncologists on treating the patient, with perhaps a reduced consideration of prophylaxis and prevention. A third factor may be a concern about thrombocytopenia. However, in our study of prophylaxis rates in US medical centers, we excluded patients who had thrombocytopenia, yet rates of prophylaxis were still low. Nothing in the literature indicates that anticoagulants cannot be used in patients with platelet counts of 50,000 to 150,000 cells/µL or higher, so this suggests that we need to do more education.

Dr. Jaffer: Dr. Brotman, can you tell us more about how clinicians in practice should use prophylaxis in their neurosurgery patients, such as those undergoing craniotomy or spine surgery for cancer? What is the safest and most efficacious way to prevent DVT in these patients?

Dr. Brotman: First, it’s important to recognize that some sort of prophylaxis needs to be used. Neurosurgery patients are at an extremely high risk for thromboembolic events, and such events are often fatal in these patients. Having said that, the jury is still out on whether the prophylaxis in these patients should be compression devices or anticoagulation. This gives physicians some latitude in their decisions. They can decide not to use pharmacologic prophylaxis so long as they use pneumatic compression devices consistently, perhaps even starting during the operation and certainly throughout hospitalization when the patient is immobilized.

Certainly, the concerns about using full-dose anticoagulation in the immediate postoperative setting in neurosurgery patients are valid. Yet these patients are at very high risk for thromboembolic events, and if we take too cautious an approach to prophylaxis in the immediate perioperative setting, more patients are going to have thromboembolic events, at which point management decisions become much more difficult. The risk of intracranial bleeding with anticoagulation to treat a patient who develops a DVT at postoperative day 10 will certainly be higher than it would have been with lower-dose perioperative prophylactic anticoagulation. Plus, if you put in a filter at that point, the outcomes tend to be poor. Therefore, I believe there is some degree of risk that we should be willing to take with regard to perioperative bleeding, even in neurosurgery patients.

Dr. McKean: I’d like to make a point about combination prophylaxis. At many institutions, compression stockings and sequential compression devices are used preoperatively and intraoperatively, and then pharmacologic prophylaxis—for example, twice-daily UFH—is used postoperatively. There is concern that these patients are hypercoagulable, and most clinicians believe that mechanical prophylaxis alone, even with sequential compression devices plus compression stockings, is not aggressive enough in these high-risk patients.

Dr. Jaffer: Dr. Spyropoulos, what is the optimal duration of pharmacologic prophylaxis for cancer surgery patients?

Dr. Spyropoulos: First let’s consider in-hospital prophylaxis. The supportive data for in-hospital prophylaxis are strong, and the duration of therapy used in the major in-hospital prophylaxis trials was 7 to 10 days. With regard to extended prophylaxis, we have at least two moderately sized randomized controlled trials, ENOXACAN II²³ and the substudy of FAME,²⁴ that demonstrated that extending prophylaxis with LMWH at doses of 3,400 U once daily (5,000 IU of dalteparin; 40 mg of enoxaparin) reduces VTE risk at postoperative day 30. Also, recent data from the @RISTOS registry show that in cancer surgery patients, especially those having abdominal or pelvic procedures, the leading cause of 30-day mortality was VTE.⁸ This registry also shows that despite prophylaxis, the rate of symptomatic VTE can be as high as 2%, with the rate of fatal VTE approaching 1%. Thus, in cancer patients undergoing abdominal or pelvic surgery, physicians should strongly consider prophylaxis of up to 30 days’ duration.

Dr. Jaffer: One striking finding from the @RISTOS registry was that 40% of VTE events in these cancer surgery patients occurred after postoperative day 21. This really underscores the need to consider prophylaxis for at least 4 weeks in these patients in real-world practice.

Dr. Brotman: The other striking finding from that registry was that the in-hospital prophylaxis rate was quite high, about 80%, and the rate of extended prophylaxis approached 35%. These are rates that are rarely achieved in clinical practice. Yet despite these high levels of prophylaxis, patients in this registry still had a high incidence of morbidity and mortality from VTE. This suggests that we need to improve our out-of-hospital VTE prevention paradigms.

Dr. Jaffer: Dr. Deitelzweig, oncologists and internists are often unsure about whether their ambulatory cancer patients who are receiving chemotherapy should be on any form of prophylaxis. What is your opinion?

Dr. Deitelzweig: That question comes up regularly because these patients are encountered across many medical specialties. At this point, all of the large organizations, including ASCO and NCCN, are advocating that prophylaxis is not indicated for such patients.

Nearly half of orthopedic surgery patients do not receive appropriate prophylaxis for venous thromboembolism (VTE), as defined by American College of Chest Physicians (ACCP) consensus guidelines, according to a recent analysis of a nationwide database of hospital admissions.¹ Even in teaching hospitals, compliance with consensus guidelines for thromboprophylaxis is suboptimal. In a study of adherence to the ACCP guidelines for VTE prevention among 1,907 surgical patients at 10 teaching hospitals, only 45.2% of hip fracture patients received optimal VTE prophylaxis.² Rates of optimal prophylaxis were higher among patients undergoing hip arthroplasty and knee arthroplasty—84.3% and 75.9%, respectively—but were still in need of improvement.²

GROWING INTEREST IN POSTOPERATIVE VTE PROPHYLAXIS AS A QUALITY INDICATOR

As noted in the introductory article in this supplement, the Joint Commission on Accreditation of Healthcare Organizations has taken notice of these shortcomings and has proposed national consensus standards for VTE prevention and treatment.³ Among its proposed standards are two related to risk assessment and prophylaxis: whether risk assessment/prophylaxis is ordered within 24 hours of hospital admission and within 24 hours of transfer to the intensive care unit.

Other quality-monitoring initiatives are focused specifically on VTE in the surgical population. The Surgical Care Improvement Project (SCIP) has approved two quality measures with respect to VTE prevention: (1) the proportion of surgical patients for whom recommended VTE prophylaxis is ordered, and (2) the proportion of patients who receive appropriate VTE prophylaxis (based on ACCP guideline recommendations) within 24 hours before or after surgery.⁴

In the future, two other VTE-related quality measures from SCIP may be implemented by the Centers for Medicare and Medicaid Services: (1) how often intra- or postoperative pulmonary embolism (PE) is diagnosed during the index hospitalization and within 30 days of surgery, and (2) how often intra- or postoperative deep vein thrombosis (DVT) is diagnosed during the index hospitalization and within 30 days of surgery.⁵

VTE RISK IN ORTHOPEDIC SURGERY

Surgical patients can be stratified into four VTE risk levels—low, moderate, high, and highest—based on age, surgery type, surgery duration, duration of immobilization, and other risk factors.⁶ For patients undergoing orthopedic surgery, these levels may be defined according to the following patient and surgical characteristics:

• Low risk—surgery duration of less than 30 minutes, age less than 40 years, repair of small fractures
• Moderate risk—age of 40 to 60 years, arthroscopy or repair of lower leg fractures, postoperative plaster cast
• High risk—age greater than 60 years, or age 40 to 60 years with additional VTE risk factors, or immobilization for greater than 4 days
• Highest risk—hip or knee arthroplasty, hip fracture repair, repair of open lower leg fractures, major trauma or spinal cord injury, or multiple risk factors for VTE (age > 40 years, prior VTE, cancer, or hypercoagulable state).

For patients in the low-risk category, no specific prophylaxis is indicated beyond early and aggressive ambulation.⁶ For those in all other risk categories, prophylaxis with pharmacologic anticoagulant agents and/or mechanical devices is indicated, as reviewed below.

All major orthopedic procedures confer highest risk level

Notably, the “highest risk” category includes any patient undergoing hip or knee arthroplasty or hip fracture repair. Among orthopedic surgery patients in this highest-risk category, rates of VTE events in the absence of prophylaxis are as follows:⁶

• Calf DVT, 40% to 80%
• Proximal DVT, 10% to 20%
• Clinical PE, 4% to 10%
• Fatal PE, 0.2% to 5%.

Hip replacement poses greater risk than knee replacement

Within this overall highest-risk category, thromboembolic risk in the absence of prophylaxis differs among procedures. Although patients undergoing hip replacement and those undergoing knee replacement have similar rates of DVT of any type,^6,7 hip replacement is associated with higher rates of the more clinically important events, specifically proximal DVT and PE. In the absence of prophylaxis, proximal DVT occurs in 23% to 36% of hip replacement patients as opposed to 9% to 20% of knee replacement patients; similarly, PE occurs in 0.7% to 30% of hip replacement patients as compared with 1.8% to 7.0% of knee replacement patients.^6,7

What about bleeding risk?

For many orthopedic surgeons, the risk of bleeding as a result of anticoagulant prophylaxis of VTE looms larger than the risk of VTE itself. This is likely because bleeding, when it does occur, is likely to occur more acutely than VTE does and may directly compromise the result of the operation. For this reason, orthopedic surgeons may be more likely to actually witness bleeding events than VTE events (especially fatal PEs) while their patients are still under their care, leading to a misperception of the relative risks of anticoagulation-related bleeding and thromboembolism.

In reality, rates of major bleeding with pharmacologic prophylaxis of VTE are a tiny fraction of the above-listed rates of VTE events in the absence of prophylaxis in patients undergoing major orthopedic surgery. Reported 30-day rates of major bleeding in patients receiving VTE prophylaxis with heparins range from 0.2% to 1.7%; these rates barely differ from the rates among placebo recipients in the same VTE prophylaxis trials, which range from 0.2% to 1.5%^.8,9 Additionally, within the continuum of risk of major bleeding from various medical interventions, VTE prophylaxis with heparins is one of the lowest-risk interventions, posing far less risk than, for example, the use of warfarin in ischemic stroke patients or in patients older than 75 years.

As reviewed in the introductory article of this supplement, the arsenal of anticoagulants for use in VTE prophylaxis includes low-dose unfractionated heparin (UFH), low-molecular-weight heparin (LMWH) agents such as dalteparin and enoxaparin, and the factor Xa inhibitor fondaparinux. A few additional comments about these and other anticoagulant options is warranted in the specific context of orthopedic surgery.

Fondaparinux. Because most of its formal US indications are for use as VTE prophylaxis in major orthopedic surgery—including hip replacement, knee replacement, and hip fracture repair—fondaparinux has been studied more widely in orthopedic surgery patients than in the other populations reviewed earlier in this supplement. Nevertheless, its use even in these settings has remained somewhat limited. This may be because of concerns over possible increased bleeding risk relative to some other anticoagulants. Because of bleeding risk, fondaparinux is contraindicated in patients who weigh less than 50 kg, and its package insert recommends caution when it is used in the elderly due to an increased risk of bleeding in patients aged 65 or older. Additionally, the Pentasaccharide in Major Knee Surgery (PENTAMAKS) study found fondaparinux to be associated with a significantly higher incidence of major bleeding compared with enoxaparin (2.1% vs 0.2%; P = .006) in major knee surgery, although it was superior to enoxaparin in preventing VTE.¹⁰ Other possible reasons for slow adoption of fondaparinux include its long half-life, which results in a sustained antithrombotic effect, its lack of easy reversibility, and a contraindication in patients with renal insufficiency.¹¹

Limited role for UFH. Low-dose UFH has a more limited role in orthopedic surgery than in other settings requiring VTE prophylaxis, as current ACCP guidelines for VTE prevention recognize it only as a possible option in hip fracture surgery and state that it is not to be considered as sole prophylaxis in patients undergoing hip or knee replacement.⁶

Warfarin. Although not indicated for use in other VTE prophylaxis settings, the vitamin K antagonist warfarin is recommended as an option for all three major orthopedic surgery indications—knee replacement, hip replacement, and hip fracture repair.⁶

The key to effective prophylaxis with warfarin is achieving the appropriate intensity of anticoagulation. In two separate analyses, Hylek et al demonstrated a balance between safety and efficacy with warfarin therapy targeted to an international normalized ratio (INR) of 2.0 to 3.0.^12,13 An INR greater than 4.0 greatly increased the risk of intracranial hemorrhage, whereas thrombosis was not effectively prevented with an INR less than 2.0.^12,13 This latter point should be stressed to orthopedic surgeons, who sometimes aim for INR values below 2.0.

Although anticoagulation clinics are superior to usual care at maintaining the INR within the window of 2.0 to 3.0, only about one-third of patients nationally who take warfarin receive care in such clinics.¹⁴ Even with optimal care in anticoagulation clinics, some patients will still receive subtherapeutic or supertherapeutic levels of warfarin, which is one of this agent’s limitations.

Aspirin not recommended as sole agent. Although aspirin is still used as thromboprophylaxis in orthopedic surgery patients, current ACCP guidelines recommend against its use as the sole means of VTE prophylaxis in any patient group.⁶ The limitations of the evidence for aspirin in this setting are illustrated by the Pulmonary Embolism Prevention study, a multicenter randomized trial in patients undergoing hip fracture (n = 13,356) or hip/knee replacement (n = 4,088).¹⁵ Patients received aspirin 160 mg/day or placebo for 5 weeks, starting preoperatively, and were evaluated for outcomes at day 35. Among the hip fracture patients, the rate of symptomatic DVT was lower in the aspirin group than in the placebo group (1.0% vs 1.5%; P = .03), as was the rate of PE (0.7% vs 1.2%, respectively; P = .002), but there was no significant difference in outcomes between the groups among the patients undergoing hip or knee replacement. Notably, 40% of patients in the study also received UFH or LMWH. Further confounding the results, some patients received nonpharmacologic VTE prophylaxis modalities, and others received nonsteroidal anti-inflammatory drugs other than aspirin.

Heparin-induced thrombocytopenia. As noted earlier in this supplement, the incidence of heparin-induced thrombocytopenia (HIT) is markedly higher in patients who receive UFH than in those who receive LMWH. This difference in frequency, which constitutes about a sixfold to eightfold differential, is due to the relationship between standard heparin and platelet factor IV, which can induce formation of IgG antibodies.¹⁶ A 50% or greater reduction in platelet count in heparin recipients should prompt consideration of HIT.

Oral direct thrombin inhibitors. Although the oral direct thrombin inhibitor ximelagatran was rejected for approval by the US Food and Drug Administration (FDA) and recently withdrawn from the market world­wide as a result of hepatic risks, other oral direct thrombin inhibitors are in phase 3 studies for use in orthopedic surgery and may be commercially available options for postoperative VTE prophylaxis before long.

GUIDELINES FOR VTE PROPHYLAXIS IN ORTHOPEDIC SURGERY

The ACCP guidelines referred to throughout this article are widely recognized as a practice standard for VTE prevention and treatment, and have been regularly updated throughout recent decades. The most recent version, issued in 2004, is formally known as the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy.⁶ Key orthopedic surgery-related recommendations and notable changes from the previous version of the guidelines, issued in 2001, are outlined below, along with pertinent supportive or illustrative studies.

Hip replacement surgery

In a change from the previous guidelines, the Seventh ACCP Conference recommends extended prophylaxis, for up to 28 to 35 days after surgery, for patients undergoing hip replacement or hip fracture surgery. For hip replacement surgery, this is a Grade 1A recommendation for prophylaxis with either LMWH or warfarin and a Grade 1C+ recommendation (“no RCTs but strong RCT results can be unequivocally extrapolated, or overwhelming evidence from observational studies”) for prophylaxis with fondaparinux.⁶

Reprinted, with permission, from Annals of Internal Medicine (Hull et al, 2001).17

The same three anticoagulant options that received Grade 1A recommendations for patients undergoing total hip replacement—LMWH, fondaparinux, and adjusted-dose warfarin—are also given Grade 1A recommendations as routine thromboprophylaxis in patients undergoing elective knee replacement (see Table 1 for dosing). In addition, optimal use of intermittent pneumatic compression devices is recommended as an alternative option to anticoagulant prophylaxis in these patients (Grade 1B, indicating a “strong recommendation” based on RCTs with important limitations). Use of UFH as the sole agent for prophylaxis is recommended against.⁶

For both hip and knee replacement surgery, the Seventh ACCP Conference does not endorse superiority of any one of its three recommended prophylaxis options—LMWH, fondaparinux, and adjusted-dose warfarin—over the other two. However, at least four large randomized trials have directly compared LMWH and adjusted-dose warfarin in the setting of arthroplasty—two in total hip replacement surgery^19,20 and two in total knee replacement surgery.^21,22 Each of these four studies found LMWH to be significantly more effective than warfarin in preventing VTE. In three of the four trials, there was no significant difference between the therapies in rates of major bleeding.^19,21,22 In the remaining trial, which was conducted in hip replacement surgery patients and compared postoperative warfarin with dalteparin initiated either immediately before or early after surgery, patients who received preoperative dalteparin initiation (but not those who received postoperative dalteparin initiation) had an increased rate of major bleeding compared with warfarin recipients (P = .01).²⁰

Hip fracture surgery

The supportive evidence for anticoagulant prophylaxis in hip fracture surgery is less robust than that in hip and knee replacement surgery. As a result, only fondaparinux has a Grade 1A recommendation as routine prophylaxis in patients undergoing hip fracture surgery. Options with less definitive recommendations are LMWH (Grade 1C+), low-dose UFH (Grade 1B), and adjusted-dose warfarin (Grade 2B, indicating a “weak recommendation” based on RCTs with important limitations) (see Table 1 for dosing of all agents).⁶

These differing recommendations are supported by the double-blind Pentasaccharide in Hip Fracture Surgery Study (PENTHIFRA) of 1,711 consecutive patients undergoing surgery for hip fracture repair.²³ Patients were randomized to at least 5 days of fondaparinux 2.5 mg once daily, initiated postoperatively, or enoxaparin 40 mg once daily, initiated preoperatively. The incidence of DVT or PE by postoperative day 11 was 8.3% in the fondaparinux arm versus 19.1% in the enoxaparin arm, a statistically significant difference (P < .001) in favor of fondaparinux. There were no differences between the groups in rates of death or clinically relevant bleeding.

As noted above, the newly added recommendation in the Seventh ACCP Conference for extended prophylaxis, for up to 28 to 35 days after surgery, applies to patients undergoing hip fracture surgery as well as those undergoing hip replacement surgery. In the setting of hip fracture repair, extended prophylaxis is a Grade 1A recommendation with the use of fondaparinux and a Grade 1C+ recommendation with the use of either LMWH or adjusted-dose warfarin.⁶

Lower extremity fractures and trauma

Although lower extremity fractures are very common, the risk of DVT has been poorly studied in this setting. For patients with isolated lower extremity fractures, the Seventh ACCP Conference recommends that clinicians not use thromboprophylaxis routinely (Grade 2A, indicating an “intermediate-strength recommendation” based on RCTs without important limitations).⁶

Trauma patients, in contrast, are well recognized as being at very high risk for DVT and PE. The Seventh ACCP Conference gives a Grade 1A recommendation to thromboprophylaxis for all trauma patients who have at least one risk factor for VTE. LMWH is recommended (Grade 1A) as the agent of choice for this purpose, provided there are no contraindications to its use, and should be administered as soon as safely possible. Mechanical modalities are reserved for trauma patients with active bleeding or high risk for hemorrhage (Grade 1B). The guidelines recommend against use of inferior vena cava (IVC) filters as primary thromboprophylaxis in trauma patients (Grade 1C, indicating an “intermediate-strength recommendation” based on observational studies).⁶

Use of ultrasonography

Duplex ultrasonographic screening is recommended in orthopedic trauma patients who are at high risk for VTE and have received suboptimal or no prophylaxis (Grade 1C). In contrast, the Seventh ACCP Conference recommends against routine use of duplex ultrasonography to screen for VTE at hospital discharge in asymptomatic patients following major orthopedic surgery (Grade 1A).⁶

Knee arthroscopy

Arthroscopic knee procedures are increasing in frequency and raise the specter of a potential role for thromboprophylaxis. However, the clinical diagnosis of DVT is unreliable, and even diagnosis by ultrasonography is unreliable following knee arthroscopy, as interpreting scans of veins below the knee is challenging in this setting.²⁴

The Seventh ACCP Conference recommends that clinicians not use routine thromboprophylaxis, other than early mobilization, for patients who undergo knee arthroscopy (Grade 2B). However, for arthroscopy patients who have inherent risk factors for VTE or who undergo a prolonged or complicated arthroscopy procedure, thromboprophylaxis with LMWH is suggested (Grade 2B).⁶

RECOMMENDED APPROACH TO VTE PROPHYLAXIS IN ORTHOPEDIC SURGERY

Drawing on the ACCP guidelines and the evidence reviewed above, we have outlined our evidence-based recommendations for pharmacologic VTE prophylaxis in patients undergoing orthopedic surgery, as presented in Table 1. All patients undergoing major orthopedic surgical procedures (ie, procedures other than arthroscopy) should routinely receive anticoagulant prophylaxis unless they have contraindications to anticoagulation. Recommended agents and their duration of use vary according to the type of surgery, as detailed in Table 1.

Extended-duration prophylaxis is recommended for patients undergoing total hip replacement and hip fracture surgery. Aspirin is not recommended as the sole agent for prophylaxis in any orthopedic surgery setting.

Importance of a postoperative prophylaxis protocol

In addition to these broad pharmacologic recommendations, it is important that a postoperative VTE prophylaxis protocol be in place at all hospitals.

At the Ochsner Medical Center in New Orleans, where one of us (S.B.D.) practices, postoperative orders include antithrombotic therapy for surgical patients, starting with placement of thigh-high antiembolism stockings on both legs on the day of surgery for patients undergoing hip replacement and on postoperative day 1 in those undergoing knee replacement. Plantar pneumatic compression devices are applied to both legs in the recovery room and kept on except when the patient is walking. The hospitalist team dictates further anticoagulation orders. If extended prophylaxis is prescribed, the discharge planner sets up drug delivery and reimbursement, provides a LMWH discharge kit, and teaches the patient to self-inject. If there is concern about increasing swelling at the surgical site while anticoagulant therapy continues, the protocol calls for prompt notification of the responsible physician. To minimize the risk that spinal or epidural hematomas will develop, all agents that increase bleeding propensity should be recognized and ordered accordingly.

VTE in patients undergoing major orthopedic surgery is a serious health problem that is highly preventable, yet VTE prophylaxis remains underused in this patient population. Despite the availability of practice guidelines for VTE prevention in the orthopedic surgery setting, recommendations are not widely implemented in clinical practice. Recommended prophylactic options differ somewhat among various orthopedic procedures, and the supportive evidence differs for various anticoagulant options.

DISCUSSION: ADDITIONAL PERSPECTIVES FROM THE AUTHORS

Dr. Jaffer: The ACCP recommends against the routine use of aspirin as primary prophylaxis against VTE in major orthopedic surgery, yet orthopedic surgeons across the country still continue to use aspirin in this setting. What are your thoughts on this, Dr. McKean?

Dr. McKean: We agree with the ACCP’s recommendation against aspirin as primary VTE prophylaxis in orthopedic patients. The percentage of US knee arthroplasty patients who develop VTE after receiving no prophylaxis at all is roughly 64%; this percentage declines only slightly (to 56%) for knee arthroplasty patients who receive prophylaxis with aspirin.25 Since we clearly want to reduce VTE risk as much as possible, I would not use aspirin alone. I would use it only if the patient were already on aspirin, but then I would add either LMWH or fondaparinux.

Dr. Jaffer: Warfarin is another agent that is widely used for prophylaxis in major orthopedic surgery. In fact, the large registries of VTE prevention in major orthopedic surgery suggest that the use of warfarin may be slightly higher than the use of LMWH. If clinicians choose to use warfarin in their practice, what are your recommendations, Dr. Deitelzweig?

Dr. Deitelzweig: As primary prophylaxis for orthopedic surgery patients, warfarin must be dosed to achieve an INR of 2.0 to 3.0; the need for a value in this range is unequivocal. This is a challenging target to attain in the hospital setting.

Dr. Brotman: A study I was involved with a few years ago suggested that warfarin may be inadequate for VTE prevention in the first few days after orthopedic surgery.²⁶ Orthopedic surgeons at the Cleveland Clinic, where I was practicing at the time, routinely used systematic ultrasonography to assess for thrombosis on postoperative day 2 or 3 following hip or knee arthroplasty, so we conducted a secondary analysis of a case-control study in these ultrasonographically screened arthroplasty patients to assess rates of early VTE and look for any associations with the type of prophylaxis used. We found that there was about a tenfold increase in the risk of VTE, both distal and proximal, on postoperative day 2 or 3 among patients who received warfarin compared with those who received LMWH. We concluded that warfarin’s delayed antithrombotic effects may not provide sufficient VTE prophylaxis in the immediate postoperative setting.²⁶

Dr. Deitelzweig: That’s a good point. Although it’s important to achieve a therapeutic level of warfarin, we now have evidence that it takes some time to achieve that level, and in the interim, bad things can happen to patients.

Dr. Jaffer: Orthopedic surgery encompasses several types of procedures. Dr. Amin, which specific orthopedic surgery patients stand to benefit from extended prophylaxis, and how long should extended prophylaxis last?

Dr. Amin: Major orthopedic surgery comprises hip fracture repair, total hip replacement, and total knee replacement. For hip fracture, there are strong data to support the use of extended prophylaxis with fondaparinux 2.5 mg/day, which showed about an 88% relative reduction in the risk of symptomatic VTE compared with standard-duration fondaparinux (6 to 8 days) followed by matching placebo for the extended phase.²⁷ The total duration of fondaparinux therapy in the extended-duration arm was 4 to 5 weeks.

Likewise, data support extended prophylaxis in hip arthroplasty patients, for whom the recommended duration is also 4 to 5 weeks. The systematic review by Hull et al¹⁷ demonstrated a 0.41 relative risk of DVT with extended-duration LMWH prophylaxis versus placebo in hip replacement patients (Figure 1), which was a highly statistically significant result.

In contrast, we do not yet have good data to support extended prophylaxis for patients undergoing total knee replacement, which is a bit surprising. In this setting, prophylaxis is recommended for 7 to 14 days but not beyond that.

Dr. Jaffer: Arthroscopy is probably the most common orthopedic procedure performed in the United States today. Dr. Brotman, what is the role of prophylaxis in patients undergoing arthroscopy?

Dr. Brotman: Minor surgery such as arthroscopy can typically be performed safely without routine prophylaxis, other than having the patient ambulate as soon as possible after the procedure. There may be exceptions to this rule, however. I believe that there is potentially a role for pharmacologic prophylaxis in arthroscopy patients who have major risk factors for VTE, such as a personal history of VTE, or who are not expected to become mobile again in a normal rapid fashion after the operation, but prophylaxis has not been studied systematically in such patients.

Dr. Jaffer: Dr. Spyropoulos, there are several new anticoagulants in the pipeline, specifically agents such as the oral direct factor Xa inhibitors and the direct thrombin inhibitors. What do recent clinical trials suggest with regard to the efficacy of these two drug classes for thromboprophylaxis in major orthopedic surgery?

Dr. Spyropoulos: The agents with the most available data are the oral direct factor Xa inhibitors apixaban and rivaroxaban and the oral direct thrombin inhibitor dabigatran. For prophylaxis in orthopedic surgery populations, phase 2 studies have been completed for apixaban and phase 3 trials have been completed for rivaroxaban and dabigatran.

It appears that the factor Xa inhibitors, apixaban and rivaroxaban, are efficacious in comparison with both adjusted-dose warfarin and LMWH, which is the gold standard for this group of patients.^28,29 So these indeed appear to be promising agents. Rivaroxaban has been submitted to European regulatory agencies for approval for the prevention of VTE in patients undergoing major orthopedic surgery, and its developer plans to submit it to the FDA in 2008 for a similar indication in the United States.

The data are more equivocal with dabigatran. There have been several positive phase 3 studies in orthopedic surgery comparing two dabigatran dosing schemes, 150 and 220 mg once daily, with the European regimen of enoxaparin (40 mg once daily),³⁰ but a recent study that compared these doses with the North American enoxaparin regimen (30 mg twice daily) failed to meet the criteria for noninferiority.³¹ Further clinical trial development is necessary for dabigatran, although in January 2008 the European Medicines Agency recommended its marketing approval for thromboprophylaxis in patients undergoing orthopedic procedures.³²

I believe that in the next 3 to 5 years our armamentarium will see the addition of at least one, if not more, of these new agents that offer the promise of oral anticoagulation with highly predictable pharmacokinetics and pharmacodynamics and no need for monitoring.

Nearly half of orthopedic surgery patients do not receive appropriate prophylaxis for venous thromboembolism (VTE), as defined by American College of Chest Physicians (ACCP) consensus guidelines, according to a recent analysis of a nationwide database of hospital admissions.¹ Even in teaching hospitals, compliance with consensus guidelines for thromboprophylaxis is suboptimal. In a study of adherence to the ACCP guidelines for VTE prevention among 1,907 surgical patients at 10 teaching hospitals, only 45.2% of hip fracture patients received optimal VTE prophylaxis.² Rates of optimal prophylaxis were higher among patients undergoing hip arthroplasty and knee arthroplasty—84.3% and 75.9%, respectively—but were still in need of improvement.²

GROWING INTEREST IN POSTOPERATIVE VTE PROPHYLAXIS AS A QUALITY INDICATOR

As noted in the introductory article in this supplement, the Joint Commission on Accreditation of Healthcare Organizations has taken notice of these shortcomings and has proposed national consensus standards for VTE prevention and treatment.³ Among its proposed standards are two related to risk assessment and prophylaxis: whether risk assessment/prophylaxis is ordered within 24 hours of hospital admission and within 24 hours of transfer to the intensive care unit.

Other quality-monitoring initiatives are focused specifically on VTE in the surgical population. The Surgical Care Improvement Project (SCIP) has approved two quality measures with respect to VTE prevention: (1) the proportion of surgical patients for whom recommended VTE prophylaxis is ordered, and (2) the proportion of patients who receive appropriate VTE prophylaxis (based on ACCP guideline recommendations) within 24 hours before or after surgery.⁴

In the future, two other VTE-related quality measures from SCIP may be implemented by the Centers for Medicare and Medicaid Services: (1) how often intra- or postoperative pulmonary embolism (PE) is diagnosed during the index hospitalization and within 30 days of surgery, and (2) how often intra- or postoperative deep vein thrombosis (DVT) is diagnosed during the index hospitalization and within 30 days of surgery.⁵

VTE RISK IN ORTHOPEDIC SURGERY

Surgical patients can be stratified into four VTE risk levels—low, moderate, high, and highest—based on age, surgery type, surgery duration, duration of immobilization, and other risk factors.⁶ For patients undergoing orthopedic surgery, these levels may be defined according to the following patient and surgical characteristics:

• Low risk—surgery duration of less than 30 minutes, age less than 40 years, repair of small fractures
• Moderate risk—age of 40 to 60 years, arthroscopy or repair of lower leg fractures, postoperative plaster cast
• High risk—age greater than 60 years, or age 40 to 60 years with additional VTE risk factors, or immobilization for greater than 4 days
• Highest risk—hip or knee arthroplasty, hip fracture repair, repair of open lower leg fractures, major trauma or spinal cord injury, or multiple risk factors for VTE (age > 40 years, prior VTE, cancer, or hypercoagulable state).

For patients in the low-risk category, no specific prophylaxis is indicated beyond early and aggressive ambulation.⁶ For those in all other risk categories, prophylaxis with pharmacologic anticoagulant agents and/or mechanical devices is indicated, as reviewed below.

All major orthopedic procedures confer highest risk level

Notably, the “highest risk” category includes any patient undergoing hip or knee arthroplasty or hip fracture repair. Among orthopedic surgery patients in this highest-risk category, rates of VTE events in the absence of prophylaxis are as follows:⁶

• Calf DVT, 40% to 80%
• Proximal DVT, 10% to 20%
• Clinical PE, 4% to 10%
• Fatal PE, 0.2% to 5%.

Hip replacement poses greater risk than knee replacement

Within this overall highest-risk category, thromboembolic risk in the absence of prophylaxis differs among procedures. Although patients undergoing hip replacement and those undergoing knee replacement have similar rates of DVT of any type,^6,7 hip replacement is associated with higher rates of the more clinically important events, specifically proximal DVT and PE. In the absence of prophylaxis, proximal DVT occurs in 23% to 36% of hip replacement patients as opposed to 9% to 20% of knee replacement patients; similarly, PE occurs in 0.7% to 30% of hip replacement patients as compared with 1.8% to 7.0% of knee replacement patients.^6,7

What about bleeding risk?

For many orthopedic surgeons, the risk of bleeding as a result of anticoagulant prophylaxis of VTE looms larger than the risk of VTE itself. This is likely because bleeding, when it does occur, is likely to occur more acutely than VTE does and may directly compromise the result of the operation. For this reason, orthopedic surgeons may be more likely to actually witness bleeding events than VTE events (especially fatal PEs) while their patients are still under their care, leading to a misperception of the relative risks of anticoagulation-related bleeding and thromboembolism.

In reality, rates of major bleeding with pharmacologic prophylaxis of VTE are a tiny fraction of the above-listed rates of VTE events in the absence of prophylaxis in patients undergoing major orthopedic surgery. Reported 30-day rates of major bleeding in patients receiving VTE prophylaxis with heparins range from 0.2% to 1.7%; these rates barely differ from the rates among placebo recipients in the same VTE prophylaxis trials, which range from 0.2% to 1.5%^.8,9 Additionally, within the continuum of risk of major bleeding from various medical interventions, VTE prophylaxis with heparins is one of the lowest-risk interventions, posing far less risk than, for example, the use of warfarin in ischemic stroke patients or in patients older than 75 years.

As reviewed in the introductory article of this supplement, the arsenal of anticoagulants for use in VTE prophylaxis includes low-dose unfractionated heparin (UFH), low-molecular-weight heparin (LMWH) agents such as dalteparin and enoxaparin, and the factor Xa inhibitor fondaparinux. A few additional comments about these and other anticoagulant options is warranted in the specific context of orthopedic surgery.

Fondaparinux. Because most of its formal US indications are for use as VTE prophylaxis in major orthopedic surgery—including hip replacement, knee replacement, and hip fracture repair—fondaparinux has been studied more widely in orthopedic surgery patients than in the other populations reviewed earlier in this supplement. Nevertheless, its use even in these settings has remained somewhat limited. This may be because of concerns over possible increased bleeding risk relative to some other anticoagulants. Because of bleeding risk, fondaparinux is contraindicated in patients who weigh less than 50 kg, and its package insert recommends caution when it is used in the elderly due to an increased risk of bleeding in patients aged 65 or older. Additionally, the Pentasaccharide in Major Knee Surgery (PENTAMAKS) study found fondaparinux to be associated with a significantly higher incidence of major bleeding compared with enoxaparin (2.1% vs 0.2%; P = .006) in major knee surgery, although it was superior to enoxaparin in preventing VTE.¹⁰ Other possible reasons for slow adoption of fondaparinux include its long half-life, which results in a sustained antithrombotic effect, its lack of easy reversibility, and a contraindication in patients with renal insufficiency.¹¹

Limited role for UFH. Low-dose UFH has a more limited role in orthopedic surgery than in other settings requiring VTE prophylaxis, as current ACCP guidelines for VTE prevention recognize it only as a possible option in hip fracture surgery and state that it is not to be considered as sole prophylaxis in patients undergoing hip or knee replacement.⁶

Warfarin. Although not indicated for use in other VTE prophylaxis settings, the vitamin K antagonist warfarin is recommended as an option for all three major orthopedic surgery indications—knee replacement, hip replacement, and hip fracture repair.⁶

The key to effective prophylaxis with warfarin is achieving the appropriate intensity of anticoagulation. In two separate analyses, Hylek et al demonstrated a balance between safety and efficacy with warfarin therapy targeted to an international normalized ratio (INR) of 2.0 to 3.0.^12,13 An INR greater than 4.0 greatly increased the risk of intracranial hemorrhage, whereas thrombosis was not effectively prevented with an INR less than 2.0.^12,13 This latter point should be stressed to orthopedic surgeons, who sometimes aim for INR values below 2.0.

Although anticoagulation clinics are superior to usual care at maintaining the INR within the window of 2.0 to 3.0, only about one-third of patients nationally who take warfarin receive care in such clinics.¹⁴ Even with optimal care in anticoagulation clinics, some patients will still receive subtherapeutic or supertherapeutic levels of warfarin, which is one of this agent’s limitations.

Aspirin not recommended as sole agent. Although aspirin is still used as thromboprophylaxis in orthopedic surgery patients, current ACCP guidelines recommend against its use as the sole means of VTE prophylaxis in any patient group.⁶ The limitations of the evidence for aspirin in this setting are illustrated by the Pulmonary Embolism Prevention study, a multicenter randomized trial in patients undergoing hip fracture (n = 13,356) or hip/knee replacement (n = 4,088).¹⁵ Patients received aspirin 160 mg/day or placebo for 5 weeks, starting preoperatively, and were evaluated for outcomes at day 35. Among the hip fracture patients, the rate of symptomatic DVT was lower in the aspirin group than in the placebo group (1.0% vs 1.5%; P = .03), as was the rate of PE (0.7% vs 1.2%, respectively; P = .002), but there was no significant difference in outcomes between the groups among the patients undergoing hip or knee replacement. Notably, 40% of patients in the study also received UFH or LMWH. Further confounding the results, some patients received nonpharmacologic VTE prophylaxis modalities, and others received nonsteroidal anti-inflammatory drugs other than aspirin.

Heparin-induced thrombocytopenia. As noted earlier in this supplement, the incidence of heparin-induced thrombocytopenia (HIT) is markedly higher in patients who receive UFH than in those who receive LMWH. This difference in frequency, which constitutes about a sixfold to eightfold differential, is due to the relationship between standard heparin and platelet factor IV, which can induce formation of IgG antibodies.¹⁶ A 50% or greater reduction in platelet count in heparin recipients should prompt consideration of HIT.

Oral direct thrombin inhibitors. Although the oral direct thrombin inhibitor ximelagatran was rejected for approval by the US Food and Drug Administration (FDA) and recently withdrawn from the market world­wide as a result of hepatic risks, other oral direct thrombin inhibitors are in phase 3 studies for use in orthopedic surgery and may be commercially available options for postoperative VTE prophylaxis before long.

GUIDELINES FOR VTE PROPHYLAXIS IN ORTHOPEDIC SURGERY

The ACCP guidelines referred to throughout this article are widely recognized as a practice standard for VTE prevention and treatment, and have been regularly updated throughout recent decades. The most recent version, issued in 2004, is formally known as the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy.⁶ Key orthopedic surgery-related recommendations and notable changes from the previous version of the guidelines, issued in 2001, are outlined below, along with pertinent supportive or illustrative studies.

Hip replacement surgery

In a change from the previous guidelines, the Seventh ACCP Conference recommends extended prophylaxis, for up to 28 to 35 days after surgery, for patients undergoing hip replacement or hip fracture surgery. For hip replacement surgery, this is a Grade 1A recommendation for prophylaxis with either LMWH or warfarin and a Grade 1C+ recommendation (“no RCTs but strong RCT results can be unequivocally extrapolated, or overwhelming evidence from observational studies”) for prophylaxis with fondaparinux.⁶

Reprinted, with permission, from Annals of Internal Medicine (Hull et al, 2001).17

The same three anticoagulant options that received Grade 1A recommendations for patients undergoing total hip replacement—LMWH, fondaparinux, and adjusted-dose warfarin—are also given Grade 1A recommendations as routine thromboprophylaxis in patients undergoing elective knee replacement (see Table 1 for dosing). In addition, optimal use of intermittent pneumatic compression devices is recommended as an alternative option to anticoagulant prophylaxis in these patients (Grade 1B, indicating a “strong recommendation” based on RCTs with important limitations). Use of UFH as the sole agent for prophylaxis is recommended against.⁶

For both hip and knee replacement surgery, the Seventh ACCP Conference does not endorse superiority of any one of its three recommended prophylaxis options—LMWH, fondaparinux, and adjusted-dose warfarin—over the other two. However, at least four large randomized trials have directly compared LMWH and adjusted-dose warfarin in the setting of arthroplasty—two in total hip replacement surgery^19,20 and two in total knee replacement surgery.^21,22 Each of these four studies found LMWH to be significantly more effective than warfarin in preventing VTE. In three of the four trials, there was no significant difference between the therapies in rates of major bleeding.^19,21,22 In the remaining trial, which was conducted in hip replacement surgery patients and compared postoperative warfarin with dalteparin initiated either immediately before or early after surgery, patients who received preoperative dalteparin initiation (but not those who received postoperative dalteparin initiation) had an increased rate of major bleeding compared with warfarin recipients (P = .01).²⁰

Hip fracture surgery

The supportive evidence for anticoagulant prophylaxis in hip fracture surgery is less robust than that in hip and knee replacement surgery. As a result, only fondaparinux has a Grade 1A recommendation as routine prophylaxis in patients undergoing hip fracture surgery. Options with less definitive recommendations are LMWH (Grade 1C+), low-dose UFH (Grade 1B), and adjusted-dose warfarin (Grade 2B, indicating a “weak recommendation” based on RCTs with important limitations) (see Table 1 for dosing of all agents).⁶

These differing recommendations are supported by the double-blind Pentasaccharide in Hip Fracture Surgery Study (PENTHIFRA) of 1,711 consecutive patients undergoing surgery for hip fracture repair.²³ Patients were randomized to at least 5 days of fondaparinux 2.5 mg once daily, initiated postoperatively, or enoxaparin 40 mg once daily, initiated preoperatively. The incidence of DVT or PE by postoperative day 11 was 8.3% in the fondaparinux arm versus 19.1% in the enoxaparin arm, a statistically significant difference (P < .001) in favor of fondaparinux. There were no differences between the groups in rates of death or clinically relevant bleeding.

As noted above, the newly added recommendation in the Seventh ACCP Conference for extended prophylaxis, for up to 28 to 35 days after surgery, applies to patients undergoing hip fracture surgery as well as those undergoing hip replacement surgery. In the setting of hip fracture repair, extended prophylaxis is a Grade 1A recommendation with the use of fondaparinux and a Grade 1C+ recommendation with the use of either LMWH or adjusted-dose warfarin.⁶

Lower extremity fractures and trauma

Although lower extremity fractures are very common, the risk of DVT has been poorly studied in this setting. For patients with isolated lower extremity fractures, the Seventh ACCP Conference recommends that clinicians not use thromboprophylaxis routinely (Grade 2A, indicating an “intermediate-strength recommendation” based on RCTs without important limitations).⁶

Trauma patients, in contrast, are well recognized as being at very high risk for DVT and PE. The Seventh ACCP Conference gives a Grade 1A recommendation to thromboprophylaxis for all trauma patients who have at least one risk factor for VTE. LMWH is recommended (Grade 1A) as the agent of choice for this purpose, provided there are no contraindications to its use, and should be administered as soon as safely possible. Mechanical modalities are reserved for trauma patients with active bleeding or high risk for hemorrhage (Grade 1B). The guidelines recommend against use of inferior vena cava (IVC) filters as primary thromboprophylaxis in trauma patients (Grade 1C, indicating an “intermediate-strength recommendation” based on observational studies).⁶

Use of ultrasonography

Duplex ultrasonographic screening is recommended in orthopedic trauma patients who are at high risk for VTE and have received suboptimal or no prophylaxis (Grade 1C). In contrast, the Seventh ACCP Conference recommends against routine use of duplex ultrasonography to screen for VTE at hospital discharge in asymptomatic patients following major orthopedic surgery (Grade 1A).⁶

Knee arthroscopy

Arthroscopic knee procedures are increasing in frequency and raise the specter of a potential role for thromboprophylaxis. However, the clinical diagnosis of DVT is unreliable, and even diagnosis by ultrasonography is unreliable following knee arthroscopy, as interpreting scans of veins below the knee is challenging in this setting.²⁴

The Seventh ACCP Conference recommends that clinicians not use routine thromboprophylaxis, other than early mobilization, for patients who undergo knee arthroscopy (Grade 2B). However, for arthroscopy patients who have inherent risk factors for VTE or who undergo a prolonged or complicated arthroscopy procedure, thromboprophylaxis with LMWH is suggested (Grade 2B).⁶

RECOMMENDED APPROACH TO VTE PROPHYLAXIS IN ORTHOPEDIC SURGERY

Drawing on the ACCP guidelines and the evidence reviewed above, we have outlined our evidence-based recommendations for pharmacologic VTE prophylaxis in patients undergoing orthopedic surgery, as presented in Table 1. All patients undergoing major orthopedic surgical procedures (ie, procedures other than arthroscopy) should routinely receive anticoagulant prophylaxis unless they have contraindications to anticoagulation. Recommended agents and their duration of use vary according to the type of surgery, as detailed in Table 1.

Extended-duration prophylaxis is recommended for patients undergoing total hip replacement and hip fracture surgery. Aspirin is not recommended as the sole agent for prophylaxis in any orthopedic surgery setting.

Importance of a postoperative prophylaxis protocol

In addition to these broad pharmacologic recommendations, it is important that a postoperative VTE prophylaxis protocol be in place at all hospitals.

At the Ochsner Medical Center in New Orleans, where one of us (S.B.D.) practices, postoperative orders include antithrombotic therapy for surgical patients, starting with placement of thigh-high antiembolism stockings on both legs on the day of surgery for patients undergoing hip replacement and on postoperative day 1 in those undergoing knee replacement. Plantar pneumatic compression devices are applied to both legs in the recovery room and kept on except when the patient is walking. The hospitalist team dictates further anticoagulation orders. If extended prophylaxis is prescribed, the discharge planner sets up drug delivery and reimbursement, provides a LMWH discharge kit, and teaches the patient to self-inject. If there is concern about increasing swelling at the surgical site while anticoagulant therapy continues, the protocol calls for prompt notification of the responsible physician. To minimize the risk that spinal or epidural hematomas will develop, all agents that increase bleeding propensity should be recognized and ordered accordingly.

VTE in patients undergoing major orthopedic surgery is a serious health problem that is highly preventable, yet VTE prophylaxis remains underused in this patient population. Despite the availability of practice guidelines for VTE prevention in the orthopedic surgery setting, recommendations are not widely implemented in clinical practice. Recommended prophylactic options differ somewhat among various orthopedic procedures, and the supportive evidence differs for various anticoagulant options.

DISCUSSION: ADDITIONAL PERSPECTIVES FROM THE AUTHORS

Dr. Jaffer: The ACCP recommends against the routine use of aspirin as primary prophylaxis against VTE in major orthopedic surgery, yet orthopedic surgeons across the country still continue to use aspirin in this setting. What are your thoughts on this, Dr. McKean?

Dr. McKean: We agree with the ACCP’s recommendation against aspirin as primary VTE prophylaxis in orthopedic patients. The percentage of US knee arthroplasty patients who develop VTE after receiving no prophylaxis at all is roughly 64%; this percentage declines only slightly (to 56%) for knee arthroplasty patients who receive prophylaxis with aspirin.25 Since we clearly want to reduce VTE risk as much as possible, I would not use aspirin alone. I would use it only if the patient were already on aspirin, but then I would add either LMWH or fondaparinux.

Dr. Jaffer: Warfarin is another agent that is widely used for prophylaxis in major orthopedic surgery. In fact, the large registries of VTE prevention in major orthopedic surgery suggest that the use of warfarin may be slightly higher than the use of LMWH. If clinicians choose to use warfarin in their practice, what are your recommendations, Dr. Deitelzweig?

Dr. Deitelzweig: As primary prophylaxis for orthopedic surgery patients, warfarin must be dosed to achieve an INR of 2.0 to 3.0; the need for a value in this range is unequivocal. This is a challenging target to attain in the hospital setting.

Dr. Brotman: A study I was involved with a few years ago suggested that warfarin may be inadequate for VTE prevention in the first few days after orthopedic surgery.²⁶ Orthopedic surgeons at the Cleveland Clinic, where I was practicing at the time, routinely used systematic ultrasonography to assess for thrombosis on postoperative day 2 or 3 following hip or knee arthroplasty, so we conducted a secondary analysis of a case-control study in these ultrasonographically screened arthroplasty patients to assess rates of early VTE and look for any associations with the type of prophylaxis used. We found that there was about a tenfold increase in the risk of VTE, both distal and proximal, on postoperative day 2 or 3 among patients who received warfarin compared with those who received LMWH. We concluded that warfarin’s delayed antithrombotic effects may not provide sufficient VTE prophylaxis in the immediate postoperative setting.²⁶

Dr. Deitelzweig: That’s a good point. Although it’s important to achieve a therapeutic level of warfarin, we now have evidence that it takes some time to achieve that level, and in the interim, bad things can happen to patients.

Dr. Jaffer: Orthopedic surgery encompasses several types of procedures. Dr. Amin, which specific orthopedic surgery patients stand to benefit from extended prophylaxis, and how long should extended prophylaxis last?

Dr. Amin: Major orthopedic surgery comprises hip fracture repair, total hip replacement, and total knee replacement. For hip fracture, there are strong data to support the use of extended prophylaxis with fondaparinux 2.5 mg/day, which showed about an 88% relative reduction in the risk of symptomatic VTE compared with standard-duration fondaparinux (6 to 8 days) followed by matching placebo for the extended phase.²⁷ The total duration of fondaparinux therapy in the extended-duration arm was 4 to 5 weeks.

Likewise, data support extended prophylaxis in hip arthroplasty patients, for whom the recommended duration is also 4 to 5 weeks. The systematic review by Hull et al¹⁷ demonstrated a 0.41 relative risk of DVT with extended-duration LMWH prophylaxis versus placebo in hip replacement patients (Figure 1), which was a highly statistically significant result.

In contrast, we do not yet have good data to support extended prophylaxis for patients undergoing total knee replacement, which is a bit surprising. In this setting, prophylaxis is recommended for 7 to 14 days but not beyond that.

Dr. Jaffer: Arthroscopy is probably the most common orthopedic procedure performed in the United States today. Dr. Brotman, what is the role of prophylaxis in patients undergoing arthroscopy?

Dr. Brotman: Minor surgery such as arthroscopy can typically be performed safely without routine prophylaxis, other than having the patient ambulate as soon as possible after the procedure. There may be exceptions to this rule, however. I believe that there is potentially a role for pharmacologic prophylaxis in arthroscopy patients who have major risk factors for VTE, such as a personal history of VTE, or who are not expected to become mobile again in a normal rapid fashion after the operation, but prophylaxis has not been studied systematically in such patients.

Dr. Jaffer: Dr. Spyropoulos, there are several new anticoagulants in the pipeline, specifically agents such as the oral direct factor Xa inhibitors and the direct thrombin inhibitors. What do recent clinical trials suggest with regard to the efficacy of these two drug classes for thromboprophylaxis in major orthopedic surgery?

Dr. Spyropoulos: The agents with the most available data are the oral direct factor Xa inhibitors apixaban and rivaroxaban and the oral direct thrombin inhibitor dabigatran. For prophylaxis in orthopedic surgery populations, phase 2 studies have been completed for apixaban and phase 3 trials have been completed for rivaroxaban and dabigatran.

It appears that the factor Xa inhibitors, apixaban and rivaroxaban, are efficacious in comparison with both adjusted-dose warfarin and LMWH, which is the gold standard for this group of patients.^28,29 So these indeed appear to be promising agents. Rivaroxaban has been submitted to European regulatory agencies for approval for the prevention of VTE in patients undergoing major orthopedic surgery, and its developer plans to submit it to the FDA in 2008 for a similar indication in the United States.

The data are more equivocal with dabigatran. There have been several positive phase 3 studies in orthopedic surgery comparing two dabigatran dosing schemes, 150 and 220 mg once daily, with the European regimen of enoxaparin (40 mg once daily),³⁰ but a recent study that compared these doses with the North American enoxaparin regimen (30 mg twice daily) failed to meet the criteria for noninferiority.³¹ Further clinical trial development is necessary for dabigatran, although in January 2008 the European Medicines Agency recommended its marketing approval for thromboprophylaxis in patients undergoing orthopedic procedures.³²

I believe that in the next 3 to 5 years our armamentarium will see the addition of at least one, if not more, of these new agents that offer the promise of oral anticoagulation with highly predictable pharmacokinetics and pharmacodynamics and no need for monitoring.

Nearly half of orthopedic surgery patients do not receive appropriate prophylaxis for venous thromboembolism (VTE), as defined by American College of Chest Physicians (ACCP) consensus guidelines, according to a recent analysis of a nationwide database of hospital admissions.¹ Even in teaching hospitals, compliance with consensus guidelines for thromboprophylaxis is suboptimal. In a study of adherence to the ACCP guidelines for VTE prevention among 1,907 surgical patients at 10 teaching hospitals, only 45.2% of hip fracture patients received optimal VTE prophylaxis.² Rates of optimal prophylaxis were higher among patients undergoing hip arthroplasty and knee arthroplasty—84.3% and 75.9%, respectively—but were still in need of improvement.²

GROWING INTEREST IN POSTOPERATIVE VTE PROPHYLAXIS AS A QUALITY INDICATOR

As noted in the introductory article in this supplement, the Joint Commission on Accreditation of Healthcare Organizations has taken notice of these shortcomings and has proposed national consensus standards for VTE prevention and treatment.³ Among its proposed standards are two related to risk assessment and prophylaxis: whether risk assessment/prophylaxis is ordered within 24 hours of hospital admission and within 24 hours of transfer to the intensive care unit.

Other quality-monitoring initiatives are focused specifically on VTE in the surgical population. The Surgical Care Improvement Project (SCIP) has approved two quality measures with respect to VTE prevention: (1) the proportion of surgical patients for whom recommended VTE prophylaxis is ordered, and (2) the proportion of patients who receive appropriate VTE prophylaxis (based on ACCP guideline recommendations) within 24 hours before or after surgery.⁴

In the future, two other VTE-related quality measures from SCIP may be implemented by the Centers for Medicare and Medicaid Services: (1) how often intra- or postoperative pulmonary embolism (PE) is diagnosed during the index hospitalization and within 30 days of surgery, and (2) how often intra- or postoperative deep vein thrombosis (DVT) is diagnosed during the index hospitalization and within 30 days of surgery.⁵

VTE RISK IN ORTHOPEDIC SURGERY

Surgical patients can be stratified into four VTE risk levels—low, moderate, high, and highest—based on age, surgery type, surgery duration, duration of immobilization, and other risk factors.⁶ For patients undergoing orthopedic surgery, these levels may be defined according to the following patient and surgical characteristics:

• Low risk—surgery duration of less than 30 minutes, age less than 40 years, repair of small fractures
• Moderate risk—age of 40 to 60 years, arthroscopy or repair of lower leg fractures, postoperative plaster cast
• High risk—age greater than 60 years, or age 40 to 60 years with additional VTE risk factors, or immobilization for greater than 4 days
• Highest risk—hip or knee arthroplasty, hip fracture repair, repair of open lower leg fractures, major trauma or spinal cord injury, or multiple risk factors for VTE (age > 40 years, prior VTE, cancer, or hypercoagulable state).

For patients in the low-risk category, no specific prophylaxis is indicated beyond early and aggressive ambulation.⁶ For those in all other risk categories, prophylaxis with pharmacologic anticoagulant agents and/or mechanical devices is indicated, as reviewed below.

All major orthopedic procedures confer highest risk level

Notably, the “highest risk” category includes any patient undergoing hip or knee arthroplasty or hip fracture repair. Among orthopedic surgery patients in this highest-risk category, rates of VTE events in the absence of prophylaxis are as follows:⁶

• Calf DVT, 40% to 80%
• Proximal DVT, 10% to 20%
• Clinical PE, 4% to 10%
• Fatal PE, 0.2% to 5%.

Hip replacement poses greater risk than knee replacement

Within this overall highest-risk category, thromboembolic risk in the absence of prophylaxis differs among procedures. Although patients undergoing hip replacement and those undergoing knee replacement have similar rates of DVT of any type,^6,7 hip replacement is associated with higher rates of the more clinically important events, specifically proximal DVT and PE. In the absence of prophylaxis, proximal DVT occurs in 23% to 36% of hip replacement patients as opposed to 9% to 20% of knee replacement patients; similarly, PE occurs in 0.7% to 30% of hip replacement patients as compared with 1.8% to 7.0% of knee replacement patients.^6,7

What about bleeding risk?

For many orthopedic surgeons, the risk of bleeding as a result of anticoagulant prophylaxis of VTE looms larger than the risk of VTE itself. This is likely because bleeding, when it does occur, is likely to occur more acutely than VTE does and may directly compromise the result of the operation. For this reason, orthopedic surgeons may be more likely to actually witness bleeding events than VTE events (especially fatal PEs) while their patients are still under their care, leading to a misperception of the relative risks of anticoagulation-related bleeding and thromboembolism.

In reality, rates of major bleeding with pharmacologic prophylaxis of VTE are a tiny fraction of the above-listed rates of VTE events in the absence of prophylaxis in patients undergoing major orthopedic surgery. Reported 30-day rates of major bleeding in patients receiving VTE prophylaxis with heparins range from 0.2% to 1.7%; these rates barely differ from the rates among placebo recipients in the same VTE prophylaxis trials, which range from 0.2% to 1.5%^.8,9 Additionally, within the continuum of risk of major bleeding from various medical interventions, VTE prophylaxis with heparins is one of the lowest-risk interventions, posing far less risk than, for example, the use of warfarin in ischemic stroke patients or in patients older than 75 years.

As reviewed in the introductory article of this supplement, the arsenal of anticoagulants for use in VTE prophylaxis includes low-dose unfractionated heparin (UFH), low-molecular-weight heparin (LMWH) agents such as dalteparin and enoxaparin, and the factor Xa inhibitor fondaparinux. A few additional comments about these and other anticoagulant options is warranted in the specific context of orthopedic surgery.

Fondaparinux. Because most of its formal US indications are for use as VTE prophylaxis in major orthopedic surgery—including hip replacement, knee replacement, and hip fracture repair—fondaparinux has been studied more widely in orthopedic surgery patients than in the other populations reviewed earlier in this supplement. Nevertheless, its use even in these settings has remained somewhat limited. This may be because of concerns over possible increased bleeding risk relative to some other anticoagulants. Because of bleeding risk, fondaparinux is contraindicated in patients who weigh less than 50 kg, and its package insert recommends caution when it is used in the elderly due to an increased risk of bleeding in patients aged 65 or older. Additionally, the Pentasaccharide in Major Knee Surgery (PENTAMAKS) study found fondaparinux to be associated with a significantly higher incidence of major bleeding compared with enoxaparin (2.1% vs 0.2%; P = .006) in major knee surgery, although it was superior to enoxaparin in preventing VTE.¹⁰ Other possible reasons for slow adoption of fondaparinux include its long half-life, which results in a sustained antithrombotic effect, its lack of easy reversibility, and a contraindication in patients with renal insufficiency.¹¹

Limited role for UFH. Low-dose UFH has a more limited role in orthopedic surgery than in other settings requiring VTE prophylaxis, as current ACCP guidelines for VTE prevention recognize it only as a possible option in hip fracture surgery and state that it is not to be considered as sole prophylaxis in patients undergoing hip or knee replacement.⁶

Warfarin. Although not indicated for use in other VTE prophylaxis settings, the vitamin K antagonist warfarin is recommended as an option for all three major orthopedic surgery indications—knee replacement, hip replacement, and hip fracture repair.⁶

The key to effective prophylaxis with warfarin is achieving the appropriate intensity of anticoagulation. In two separate analyses, Hylek et al demonstrated a balance between safety and efficacy with warfarin therapy targeted to an international normalized ratio (INR) of 2.0 to 3.0.^12,13 An INR greater than 4.0 greatly increased the risk of intracranial hemorrhage, whereas thrombosis was not effectively prevented with an INR less than 2.0.^12,13 This latter point should be stressed to orthopedic surgeons, who sometimes aim for INR values below 2.0.

Although anticoagulation clinics are superior to usual care at maintaining the INR within the window of 2.0 to 3.0, only about one-third of patients nationally who take warfarin receive care in such clinics.¹⁴ Even with optimal care in anticoagulation clinics, some patients will still receive subtherapeutic or supertherapeutic levels of warfarin, which is one of this agent’s limitations.

Aspirin not recommended as sole agent. Although aspirin is still used as thromboprophylaxis in orthopedic surgery patients, current ACCP guidelines recommend against its use as the sole means of VTE prophylaxis in any patient group.⁶ The limitations of the evidence for aspirin in this setting are illustrated by the Pulmonary Embolism Prevention study, a multicenter randomized trial in patients undergoing hip fracture (n = 13,356) or hip/knee replacement (n = 4,088).¹⁵ Patients received aspirin 160 mg/day or placebo for 5 weeks, starting preoperatively, and were evaluated for outcomes at day 35. Among the hip fracture patients, the rate of symptomatic DVT was lower in the aspirin group than in the placebo group (1.0% vs 1.5%; P = .03), as was the rate of PE (0.7% vs 1.2%, respectively; P = .002), but there was no significant difference in outcomes between the groups among the patients undergoing hip or knee replacement. Notably, 40% of patients in the study also received UFH or LMWH. Further confounding the results, some patients received nonpharmacologic VTE prophylaxis modalities, and others received nonsteroidal anti-inflammatory drugs other than aspirin.

Heparin-induced thrombocytopenia. As noted earlier in this supplement, the incidence of heparin-induced thrombocytopenia (HIT) is markedly higher in patients who receive UFH than in those who receive LMWH. This difference in frequency, which constitutes about a sixfold to eightfold differential, is due to the relationship between standard heparin and platelet factor IV, which can induce formation of IgG antibodies.¹⁶ A 50% or greater reduction in platelet count in heparin recipients should prompt consideration of HIT.

Oral direct thrombin inhibitors. Although the oral direct thrombin inhibitor ximelagatran was rejected for approval by the US Food and Drug Administration (FDA) and recently withdrawn from the market world­wide as a result of hepatic risks, other oral direct thrombin inhibitors are in phase 3 studies for use in orthopedic surgery and may be commercially available options for postoperative VTE prophylaxis before long.

GUIDELINES FOR VTE PROPHYLAXIS IN ORTHOPEDIC SURGERY

The ACCP guidelines referred to throughout this article are widely recognized as a practice standard for VTE prevention and treatment, and have been regularly updated throughout recent decades. The most recent version, issued in 2004, is formally known as the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy.⁶ Key orthopedic surgery-related recommendations and notable changes from the previous version of the guidelines, issued in 2001, are outlined below, along with pertinent supportive or illustrative studies.

Hip replacement surgery

In a change from the previous guidelines, the Seventh ACCP Conference recommends extended prophylaxis, for up to 28 to 35 days after surgery, for patients undergoing hip replacement or hip fracture surgery. For hip replacement surgery, this is a Grade 1A recommendation for prophylaxis with either LMWH or warfarin and a Grade 1C+ recommendation (“no RCTs but strong RCT results can be unequivocally extrapolated, or overwhelming evidence from observational studies”) for prophylaxis with fondaparinux.⁶

Reprinted, with permission, from Annals of Internal Medicine (Hull et al, 2001).17

The same three anticoagulant options that received Grade 1A recommendations for patients undergoing total hip replacement—LMWH, fondaparinux, and adjusted-dose warfarin—are also given Grade 1A recommendations as routine thromboprophylaxis in patients undergoing elective knee replacement (see Table 1 for dosing). In addition, optimal use of intermittent pneumatic compression devices is recommended as an alternative option to anticoagulant prophylaxis in these patients (Grade 1B, indicating a “strong recommendation” based on RCTs with important limitations). Use of UFH as the sole agent for prophylaxis is recommended against.⁶

For both hip and knee replacement surgery, the Seventh ACCP Conference does not endorse superiority of any one of its three recommended prophylaxis options—LMWH, fondaparinux, and adjusted-dose warfarin—over the other two. However, at least four large randomized trials have directly compared LMWH and adjusted-dose warfarin in the setting of arthroplasty—two in total hip replacement surgery^19,20 and two in total knee replacement surgery.^21,22 Each of these four studies found LMWH to be significantly more effective than warfarin in preventing VTE. In three of the four trials, there was no significant difference between the therapies in rates of major bleeding.^19,21,22 In the remaining trial, which was conducted in hip replacement surgery patients and compared postoperative warfarin with dalteparin initiated either immediately before or early after surgery, patients who received preoperative dalteparin initiation (but not those who received postoperative dalteparin initiation) had an increased rate of major bleeding compared with warfarin recipients (P = .01).²⁰

Hip fracture surgery

The supportive evidence for anticoagulant prophylaxis in hip fracture surgery is less robust than that in hip and knee replacement surgery. As a result, only fondaparinux has a Grade 1A recommendation as routine prophylaxis in patients undergoing hip fracture surgery. Options with less definitive recommendations are LMWH (Grade 1C+), low-dose UFH (Grade 1B), and adjusted-dose warfarin (Grade 2B, indicating a “weak recommendation” based on RCTs with important limitations) (see Table 1 for dosing of all agents).⁶

These differing recommendations are supported by the double-blind Pentasaccharide in Hip Fracture Surgery Study (PENTHIFRA) of 1,711 consecutive patients undergoing surgery for hip fracture repair.²³ Patients were randomized to at least 5 days of fondaparinux 2.5 mg once daily, initiated postoperatively, or enoxaparin 40 mg once daily, initiated preoperatively. The incidence of DVT or PE by postoperative day 11 was 8.3% in the fondaparinux arm versus 19.1% in the enoxaparin arm, a statistically significant difference (P < .001) in favor of fondaparinux. There were no differences between the groups in rates of death or clinically relevant bleeding.

As noted above, the newly added recommendation in the Seventh ACCP Conference for extended prophylaxis, for up to 28 to 35 days after surgery, applies to patients undergoing hip fracture surgery as well as those undergoing hip replacement surgery. In the setting of hip fracture repair, extended prophylaxis is a Grade 1A recommendation with the use of fondaparinux and a Grade 1C+ recommendation with the use of either LMWH or adjusted-dose warfarin.⁶

Lower extremity fractures and trauma

Although lower extremity fractures are very common, the risk of DVT has been poorly studied in this setting. For patients with isolated lower extremity fractures, the Seventh ACCP Conference recommends that clinicians not use thromboprophylaxis routinely (Grade 2A, indicating an “intermediate-strength recommendation” based on RCTs without important limitations).⁶

Trauma patients, in contrast, are well recognized as being at very high risk for DVT and PE. The Seventh ACCP Conference gives a Grade 1A recommendation to thromboprophylaxis for all trauma patients who have at least one risk factor for VTE. LMWH is recommended (Grade 1A) as the agent of choice for this purpose, provided there are no contraindications to its use, and should be administered as soon as safely possible. Mechanical modalities are reserved for trauma patients with active bleeding or high risk for hemorrhage (Grade 1B). The guidelines recommend against use of inferior vena cava (IVC) filters as primary thromboprophylaxis in trauma patients (Grade 1C, indicating an “intermediate-strength recommendation” based on observational studies).⁶

Use of ultrasonography

Duplex ultrasonographic screening is recommended in orthopedic trauma patients who are at high risk for VTE and have received suboptimal or no prophylaxis (Grade 1C). In contrast, the Seventh ACCP Conference recommends against routine use of duplex ultrasonography to screen for VTE at hospital discharge in asymptomatic patients following major orthopedic surgery (Grade 1A).⁶

Knee arthroscopy

Arthroscopic knee procedures are increasing in frequency and raise the specter of a potential role for thromboprophylaxis. However, the clinical diagnosis of DVT is unreliable, and even diagnosis by ultrasonography is unreliable following knee arthroscopy, as interpreting scans of veins below the knee is challenging in this setting.²⁴

The Seventh ACCP Conference recommends that clinicians not use routine thromboprophylaxis, other than early mobilization, for patients who undergo knee arthroscopy (Grade 2B). However, for arthroscopy patients who have inherent risk factors for VTE or who undergo a prolonged or complicated arthroscopy procedure, thromboprophylaxis with LMWH is suggested (Grade 2B).⁶

RECOMMENDED APPROACH TO VTE PROPHYLAXIS IN ORTHOPEDIC SURGERY

Drawing on the ACCP guidelines and the evidence reviewed above, we have outlined our evidence-based recommendations for pharmacologic VTE prophylaxis in patients undergoing orthopedic surgery, as presented in Table 1. All patients undergoing major orthopedic surgical procedures (ie, procedures other than arthroscopy) should routinely receive anticoagulant prophylaxis unless they have contraindications to anticoagulation. Recommended agents and their duration of use vary according to the type of surgery, as detailed in Table 1.

Extended-duration prophylaxis is recommended for patients undergoing total hip replacement and hip fracture surgery. Aspirin is not recommended as the sole agent for prophylaxis in any orthopedic surgery setting.

Importance of a postoperative prophylaxis protocol

In addition to these broad pharmacologic recommendations, it is important that a postoperative VTE prophylaxis protocol be in place at all hospitals.

At the Ochsner Medical Center in New Orleans, where one of us (S.B.D.) practices, postoperative orders include antithrombotic therapy for surgical patients, starting with placement of thigh-high antiembolism stockings on both legs on the day of surgery for patients undergoing hip replacement and on postoperative day 1 in those undergoing knee replacement. Plantar pneumatic compression devices are applied to both legs in the recovery room and kept on except when the patient is walking. The hospitalist team dictates further anticoagulation orders. If extended prophylaxis is prescribed, the discharge planner sets up drug delivery and reimbursement, provides a LMWH discharge kit, and teaches the patient to self-inject. If there is concern about increasing swelling at the surgical site while anticoagulant therapy continues, the protocol calls for prompt notification of the responsible physician. To minimize the risk that spinal or epidural hematomas will develop, all agents that increase bleeding propensity should be recognized and ordered accordingly.

VTE in patients undergoing major orthopedic surgery is a serious health problem that is highly preventable, yet VTE prophylaxis remains underused in this patient population. Despite the availability of practice guidelines for VTE prevention in the orthopedic surgery setting, recommendations are not widely implemented in clinical practice. Recommended prophylactic options differ somewhat among various orthopedic procedures, and the supportive evidence differs for various anticoagulant options.

DISCUSSION: ADDITIONAL PERSPECTIVES FROM THE AUTHORS

Dr. Jaffer: The ACCP recommends against the routine use of aspirin as primary prophylaxis against VTE in major orthopedic surgery, yet orthopedic surgeons across the country still continue to use aspirin in this setting. What are your thoughts on this, Dr. McKean?

Dr. McKean: We agree with the ACCP’s recommendation against aspirin as primary VTE prophylaxis in orthopedic patients. The percentage of US knee arthroplasty patients who develop VTE after receiving no prophylaxis at all is roughly 64%; this percentage declines only slightly (to 56%) for knee arthroplasty patients who receive prophylaxis with aspirin.25 Since we clearly want to reduce VTE risk as much as possible, I would not use aspirin alone. I would use it only if the patient were already on aspirin, but then I would add either LMWH or fondaparinux.

Dr. Jaffer: Warfarin is another agent that is widely used for prophylaxis in major orthopedic surgery. In fact, the large registries of VTE prevention in major orthopedic surgery suggest that the use of warfarin may be slightly higher than the use of LMWH. If clinicians choose to use warfarin in their practice, what are your recommendations, Dr. Deitelzweig?

Dr. Deitelzweig: As primary prophylaxis for orthopedic surgery patients, warfarin must be dosed to achieve an INR of 2.0 to 3.0; the need for a value in this range is unequivocal. This is a challenging target to attain in the hospital setting.

Dr. Brotman: A study I was involved with a few years ago suggested that warfarin may be inadequate for VTE prevention in the first few days after orthopedic surgery.²⁶ Orthopedic surgeons at the Cleveland Clinic, where I was practicing at the time, routinely used systematic ultrasonography to assess for thrombosis on postoperative day 2 or 3 following hip or knee arthroplasty, so we conducted a secondary analysis of a case-control study in these ultrasonographically screened arthroplasty patients to assess rates of early VTE and look for any associations with the type of prophylaxis used. We found that there was about a tenfold increase in the risk of VTE, both distal and proximal, on postoperative day 2 or 3 among patients who received warfarin compared with those who received LMWH. We concluded that warfarin’s delayed antithrombotic effects may not provide sufficient VTE prophylaxis in the immediate postoperative setting.²⁶

Dr. Deitelzweig: That’s a good point. Although it’s important to achieve a therapeutic level of warfarin, we now have evidence that it takes some time to achieve that level, and in the interim, bad things can happen to patients.

Dr. Jaffer: Orthopedic surgery encompasses several types of procedures. Dr. Amin, which specific orthopedic surgery patients stand to benefit from extended prophylaxis, and how long should extended prophylaxis last?

Dr. Amin: Major orthopedic surgery comprises hip fracture repair, total hip replacement, and total knee replacement. For hip fracture, there are strong data to support the use of extended prophylaxis with fondaparinux 2.5 mg/day, which showed about an 88% relative reduction in the risk of symptomatic VTE compared with standard-duration fondaparinux (6 to 8 days) followed by matching placebo for the extended phase.²⁷ The total duration of fondaparinux therapy in the extended-duration arm was 4 to 5 weeks.

Likewise, data support extended prophylaxis in hip arthroplasty patients, for whom the recommended duration is also 4 to 5 weeks. The systematic review by Hull et al¹⁷ demonstrated a 0.41 relative risk of DVT with extended-duration LMWH prophylaxis versus placebo in hip replacement patients (Figure 1), which was a highly statistically significant result.

In contrast, we do not yet have good data to support extended prophylaxis for patients undergoing total knee replacement, which is a bit surprising. In this setting, prophylaxis is recommended for 7 to 14 days but not beyond that.

Dr. Jaffer: Arthroscopy is probably the most common orthopedic procedure performed in the United States today. Dr. Brotman, what is the role of prophylaxis in patients undergoing arthroscopy?

Dr. Brotman: Minor surgery such as arthroscopy can typically be performed safely without routine prophylaxis, other than having the patient ambulate as soon as possible after the procedure. There may be exceptions to this rule, however. I believe that there is potentially a role for pharmacologic prophylaxis in arthroscopy patients who have major risk factors for VTE, such as a personal history of VTE, or who are not expected to become mobile again in a normal rapid fashion after the operation, but prophylaxis has not been studied systematically in such patients.

Dr. Jaffer: Dr. Spyropoulos, there are several new anticoagulants in the pipeline, specifically agents such as the oral direct factor Xa inhibitors and the direct thrombin inhibitors. What do recent clinical trials suggest with regard to the efficacy of these two drug classes for thromboprophylaxis in major orthopedic surgery?

Dr. Spyropoulos: The agents with the most available data are the oral direct factor Xa inhibitors apixaban and rivaroxaban and the oral direct thrombin inhibitor dabigatran. For prophylaxis in orthopedic surgery populations, phase 2 studies have been completed for apixaban and phase 3 trials have been completed for rivaroxaban and dabigatran.

It appears that the factor Xa inhibitors, apixaban and rivaroxaban, are efficacious in comparison with both adjusted-dose warfarin and LMWH, which is the gold standard for this group of patients.^28,29 So these indeed appear to be promising agents. Rivaroxaban has been submitted to European regulatory agencies for approval for the prevention of VTE in patients undergoing major orthopedic surgery, and its developer plans to submit it to the FDA in 2008 for a similar indication in the United States.

The data are more equivocal with dabigatran. There have been several positive phase 3 studies in orthopedic surgery comparing two dabigatran dosing schemes, 150 and 220 mg once daily, with the European regimen of enoxaparin (40 mg once daily),³⁰ but a recent study that compared these doses with the North American enoxaparin regimen (30 mg twice daily) failed to meet the criteria for noninferiority.³¹ Further clinical trial development is necessary for dabigatran, although in January 2008 the European Medicines Agency recommended its marketing approval for thromboprophylaxis in patients undergoing orthopedic procedures.³²

I believe that in the next 3 to 5 years our armamentarium will see the addition of at least one, if not more, of these new agents that offer the promise of oral anticoagulation with highly predictable pharmacokinetics and pharmacodynamics and no need for monitoring.