HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:
Guest Access | Sign In via User Name/Password

This Article Abstract Full Text (PDF) Free Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Article Archive Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Geerts, W. H. Articles by Ray, J. G. Search for Related Content PubMed PubMed Citation Articles by Geerts, W. H. Articles by Ray, J. G.

Prevention of Venous Thromboembolism

The Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy

Correspondence to: William H. Geerts, MD, Thromboembolism Program, Sunnybrook & Women’s College Health Sciences Centre, Room D674, 2075 Bayview Ave, Toronto, ON, Canada M4N 3M5

	Abstract

TOP Abstract 1.0 Introduction Summary of Recommendations References

 
This article discusses the prevention of venous thromboembolism (VTE) and is part of the Seventh American College of Chest Physicians Conference on Antithrombotic and Thrombolytic Therapy: Evidence-Based Guidelines. Grade 1 recommendations are strong and indicate that the benefits do, or do not, outweigh risks, burden, and costs. Grade 2 suggests that individual patients’ values may lead to different choices (for a full understanding of the grading see Guyatt et al, CHEST 2004; 126:179S–187S). Among the key recommendations in this chapter are the following. We recommend against the use of aspirin alone as thromboprophylaxis for any patient group (Grade 1A). For moderate-risk general surgery patients, we recommend prophylaxis with low-dose unfractionated heparin (LDUH) (5,000 U bid) or low-molecular-weight heparin (LMWH) [ 3,400 U once daily] (both Grade 1A). For higher risk general surgery patients, we recommend thromboprophylaxis with LDUH (5,000 U tid) or LMWH (> 3,400 U daily) [both Grade 1A]. For high-risk general surgery patients with multiple risk factors, we recommend combining pharmacologic methods (LDUH three times daily or LMWH, > 3,400 U daily) with the use of graduated compression stockings and/or intermittent pneumatic compression devices (Grade 1C+). We recommend that thromboprophylaxis be used in all patients undergoing major gynecologic surgery (Grade 1A) or major, open urologic procedures, and we recommend prophylaxis with LDUH two times or three times daily (Grade 1A). For patients undergoing elective total hip or knee arthroplasty, we recommend one of the following three anticoagulant agents: LMWH, fondaparinux, or adjusted-dose vitamin K antagonist (VKA) [international normalized ratio (INR) target, 2.5; range, 2.0 to 3.0] (all Grade 1A). For patients undergoing hip fracture surgery (HFS), we recommend the routine use of fondaparinux (Grade 1A), LMWH (Grade 1C+), VKA (target INR, 2.5; range, 2.0 to 3.0) [Grade 2B], or LDUH (Grade 1B). We recommend that patients undergoing hip or knee arthroplasty, or HFS receive thromboprophylaxis for at least 10 days (Grade 1A). We recommend that all trauma patients with at least one risk factor for VTE receive thromboprophylaxis (Grade 1A). In acutely ill medical patients who have been admitted to the hospital with congestive heart failure or severe respiratory disease, or who are confined to bed and have one or more additional risk factors, we recommend prophylaxis with LDUH (Grade 1A) or LMWH (Grade 1A). We recommend, on admission to the intensive care unit, all patients be assessed for their risk of VTE. Accordingly, most patients should receive thromboprophylaxis (Grade 1A).

Key Words: aspirin • deep-vein thrombosis • fondaparinux • heparin • low-molecular-weight heparin • prophylaxis • thromboembolism • warfarin

	1.0 Introduction

TOP Abstract 1.0 Introduction Summary of Recommendations References

 
This article systematically reviews the literature related to the risks of venous thromboembolism (VTE) and its prevention. Other evidence-based reviews are also available.¹²³

1.1 Methods
This article adhered closely to the model for developing American College of Chest Physicians guidelines that is described by Schünemann et al in this Supplement.⁴ A priori criteria for inclusion of studies were applied whenever possible (Table 1 ), and always when the results of multiple trials were pooled. The number needed to treat (NNT) was used to estimate the number of patients who would need to receive a specific thromboprophylaxis regimen to prevent one additional deep-vein thrombosis (DVT), compared with patients receiving no prophylaxis or another prophylaxis regimen. The number needed to harm (NNH) was defined as the number of patients who would need to receive the thromboprophylaxis regimen to result in one additional adverse event, such as major bleeding. In formulating the final text and recommendations, we considered the comments of external reviewers (usually 5 to 10) who provided feedback on each section of this article. Although the recommendations are evidence-based, we also provide suggestions that clinicians might find useful when the evidence is weak.

Most studies of VTE and its prevention have used sensitive diagnostic tests to detect DVT. The majority of the thrombi diagnosed by these screening tests were confined to the calf, were clinically silent, and remained so without any adverse consequences.^22232425 However, approximately 10 to 20% of calf thrombi do extend to the proximal veins,^222627282930 and, particularly in patients undergoing major surgery involving the hip, isolated femoral vein DVT is common.^31323334 There is also a strong association between asymptomatic DVT and the subsequent development of symptomatic VTE.^{223536373839404142} For example, one study⁴² found that among critical care patients with asymptomatic DVT detected by screening DUS there was a significantly greater rate of PE development during their index hospitalization compared to those patients without silent DVT (11.5% vs 0%, respectively; p = 0.01). Furthermore, the in-hospital case-fatality rate of VTE is 12%,¹² and the data suggest a case-fatality rate at 1 year of 29 to 34%.¹²⁴³

While high-risk groups for VTE can be identified, it is not possible to predict which individual patients in a given risk group will develop a clinically important thromboembolic event. Furthermore, massive PE usually occurs without warning, and there is often no potential to resuscitate patients who experience this complication.¹⁵ In 70 to 80% of patients who die in the hospital of PE, this diagnosis was not even considered prior to death.^154445464748 Although the prevention of fatal PE remains the top priority for prophylaxis programs, this outcome is uncommon in most hospital groups. Furthermore, the prevention of fatal PE is not the only objective of thromboprophylaxis. The prevention of symptomatic DVT and PE are also important objectives since these outcomes are associated with considerable acute morbidity, substantial consumption of resources, and long-term sequelae of clinical and economic significance.⁵⁴⁹

The majority of symptomatic VTE associated with hospital admissions occur after hospital discharge.^41505152 When symptomatic hospital-acquired VTE is suspected, costly diagnostic testing procedures are required and, if VTE is confirmed, therapeutic anticoagulation therapy, with its potential for serious bleeding complications, should be instituted. Therefore, the failure to prevent VTE also results in delayed hospital discharge or readmission, in complications from anticoagulation therapy, in an increased risk of long-term morbidity from the post-thrombotic syndrome, and in recurrent thrombosis in the future.³⁰⁵³⁵⁴ A high proportion of venous thrombi leave residual venous abnormalities including persistent occlusion and/or venous valvular incompetence.⁵⁴⁵⁵⁵⁶ Post-thrombotic syndrome may result in chronic leg swelling, discomfort, dermatitis, and leg ulcers, reduces patient quality of life, and has considerable adverse economic effects.^57585960 These delayed consequences of inadequate prophylaxis are often overlooked.

Reliance on symptoms or signs of early DVT is an unreliable strategy to prevent clinically important thromboembolic events. The first manifestation of VTE may be fatal PE. The routine screening of patients for asymptomatic DVT is logistically difficult and is neither effective in preventing clinically important VTE nor cost-effective.^{61626364656667} Accordingly, prophylaxis against VTE remains the most appropriate strategy to reduce the sequelae discussed above.

A vast number of randomized clinical trials over the past 30 years provide irrefutable evidence that primary thromboprophylaxis reduces DVT, PE, and fatal PE.^25068697071 PE is the most common preventable cause of hospital death and is the number one strategy to improve patient safety in hospitals.¹²⁷² The Agency for Healthcare Research and Quality has published a report entitled "Making Health Care Safer: a Critical Analysis of Patient Safety Practices."⁷² This systematic review ranked 79 patient safety interventions based on the strength of the evidence supporting more widespread implementation of these procedures. The highest ranked safety practice was the "appropriate use of prophylaxis to prevent VTE in patients at risk." This recommendation was based on overwhelming evidence that thromboprophylaxis reduces adverse patient outcomes while, at the same time, decreasing overall costs.^560737475

Concerns are sometimes raised about the complications of thromboprophylaxis, especially bleeding.⁵⁰⁷⁶ However, abundant data from metaanalyses and placebo-controlled, blinded, randomized clinical trials have demonstrated little or no increase in the rates of clinically important bleeding with prophylactic doses of low-dose unfractionated heparin (LDUH), low molecular weight heparin (LMWH), or a vitamin K antagonist (VKA).^{7177787980818283} There is good evidence that appropriately used thromboprophylaxis has a desirable risk/benefit ratio and is cost-effective.^{5606173747584} Thromboprophylaxis, therefore, provides an opportunity both to improve patient outcomes and also to reduce hospital costs.

1.3 Risk factor stratification
There are two general approaches to making thromboprophylaxis decisions. One approach considers the risk of VTE in each patient, based on their individual predisposing factors and the risk associated with their current illness or procedure. Prophylaxis is then individually prescribed based on the composite risk estimate. Formal risk assessment models for DVT have been proposed to assist with this process.^{167858687888990919293} Because the approach of individual prophylaxis prescribing, based on formal risk-assessment models, has not been adequately validated and is cumbersome without the use of computer technology, it is unlikely to be used routinely by most clinicians. Furthermore, there is little formal understanding of how the various risk factors interact to determine the position of each patient along a continuous spectrum of thromboembolic risk. One simplification of this process for surgical patients involves assigning them to one of four VTE risk levels based on the type of operation (eg, minor or major), age (eg, < 40 years, 40 to 60 years, and > 60 years), and the presence of additional risk factors (eg, cancer or previous VTE) [Table 5 ]. Despite its limitations, this classification system, which was derived using prospective study data, provides both an estimate of VTE risk and related prophylaxis recommendations.

After discussing several important issues related to the interpretation of thromboprophylaxis evidence, the remainder of this article categorizes patients according to the type of hospital service that is providing care for their primary surgical or medical disorder. Within each patient category, the risks of VTE and the effective methods of prophylaxis are discussed, if they are known. For most patient groups, sufficient numbers of randomized clinical trials are available to allow strong recommendations (ie, Grade 1A or Grade 1B) to be made with regard to the benefits and risks of specific thromboprophylaxis options.

VTE is an important health-care problem, resulting in significant mortality, morbidity, and resource expenditure. Despite the continuing need for additional data, we believe that there is sufficient evidence to recommend routine thromboprophylaxis for many hospitalized patient groups. The implementation of evidence-based and thoughtful prophylaxis strategies provides benefit to patients, and should also protect their caregivers and the hospitals providing care from legal liability. We recommend that every hospital develop a formal strategy that addresses the prevention of thromboembolic complications. This should generally be in the form of a written thromboprophylaxis policy, especially for high-risk groups.

1.4 Important issues related to studies of thromboprophylaxis
The appropriate interpretation of published information about thromboprophylaxis requires the consideration of a number of important issues.

1.4.1 Limitations of DVT screening methods
Each of the methods used to screen for DVT in clinical trials has its own limitations.⁹⁵ Fibrinogen leg scanning, also called the fibrinogen uptake test (FUT), was used extensively to detect subclinical DVT in many early prophylaxis trials.⁹⁶ The test is no longer available because of concerns about the potential for viral transmission with this human blood product. Furthermore, the FUT has been shown to lack both specificity and sensitivity for the detection of DVT,^{979899100101102} and is poorly correlated with major thromboembolic events.¹⁰³ Impedance plethysmography also has been shown to have low accuracy in the screening of asymptomatic high-risk patients, and is no longer utilized.^104105106107

Contrast venography has long been the diagnostic standard in thromboprophylaxis trials¹⁰⁸ because of its high sensitivity for detecting DVT and the availability of hard-copy images for blinded study adjudication. Many pivotal, practice-changing prophylaxis trials have used venography as the primary outcome measure of efficacy. Although venography remains an important screening test for DVT, especially in evaluating the efficacy of new antithrombotic interventions, it has a number of well-recognized limitations, including the following: (1) limited availability in many medical centers; (2) questionable clinical relevance of small or distal thrombi; (3) incomplete or nondiagnostic rates of at least 20 to 40%; (4) moderate interobserver variability in its interpretation; (5) patient discomfort and risks related to the use of a contrast agent; and (6) high financial costs.^109110111112 Furthermore, because venography is not readily repeatable, it can only provide information about thrombosis at a single point in time rather than over a longer a time course during which clinically important VTE may arise.

Venous Doppler ultrasonography (DUS) is now the most universally accepted test for the diagnosis of lower extremity DVT, because it is highly accurate for symptomatic DVT, widely available, and noninvasive, and can be repeated.¹⁰⁶¹¹² At the same time, the accuracy of DUS varies among both operators and medical centers.¹¹³ While DUS has reduced sensitivity for detecting DVT in asymptomatic patients,^{106114115116117118} the accuracy of DUS appears to be improving.¹¹⁹ The lower sensitivity of DUS for detecting small and/or nonocclusive DVTs may even be considered advantageous, since such thrombi appear to be of doubtful clinical significance.¹²⁰¹²¹ The standardization of the DUS technique is critical in reducing the potential for the false-positive test results reported in some trials.¹²² As a result of recent improvements in DUS accuracy, an increasing number of clinical trials in thromboprophylaxis are utilizing ultrasound outcomes. We believe that DUS-positive proximal DVT is a clinically relevant finding because of the known association between proximal DVT and PE, and because patients with this finding generally receive anticoagulation therapy in routine practice.

Despite the limitations of each of these screening methods, and thus the possibility of error in the estimates of the absolute rates of DVT, the relative risk reductions (RRRs), derived from studies comparing two prophylaxis regimens are likely to be valid as long as systematic bias has been reduced through the concealed randomization of patients, caregivers, and outcome adjudicators to the study interventions received, and through the complete follow-up of patients.¹²³

1.4.2 Appropriate end points in clinical trials of thromboprophylaxis
Physicians differ widely in their views on the appropriate end points for studies of thromboprophylaxis.^95112124 While some believe that contrast venography should be used as the "best" test to detect all DVTs, others argue that evidence of effectiveness should be based on a proven reduction in all-cause mortality. Both of these antithetical positions clearly have limitations.

Over the years, the majority of prophylaxis trials have used DVT, detected by sensitive screening methods, as the primary efficacy outcome. While most asymptomatic DVTs are not clinically relevant, there is strong concordance between the "surrogate" outcome of asymptomatic DVT and clinically important VTE.^3436383940 In most studies, the ratio of asymptomatic DVT to symptomatic VTE ranges from 5:1 to 10:1. However, studies that employ routine screening for DVT may underestimate the true rate of symptomatic VTE or fatal PE because early screening for, and treatment of, asymptomatic DVT virtually eliminates the potential for these thrombi to progress and become symptomatic. With few exceptions, interventions that reduce asymptomatic DVT also convey similar RRRs in symptomatic VTE.^{343839405271125}

Proving a reduction in all-cause mortality or fatal PE as the objective of a thromboprophylaxis trial is problematic. Such studies require thousands of patients, and autopsy confirmation of VTE as the cause of death is increasingly difficult. Furthermore, an insistence on mortality or fatal PE as the only important outcome dismisses the significant burden of illness due to symptomatic thromboembolic events as well as the risks of anticoagulation therapy and the utilization of health-care resources when these events arise.

We (and others) have suggested^95112124 a combination of these two approaches. Phase II and some phase III clinical trials should continue to utilize sensitive imaging modalities for the detection of largely asymptomatic DVT as a means of testing the biological efficacy of a new intervention. These studies should be followed by large clinical trials that use a clinically important VTE outcome, such as the combination of symptomatic and objectively proven DVT or PE, and asymptomatic proximal DVT detected by a noninvasive test such as DUS.

1.4.3 Mechanical methods of prophylaxis
Mechanical methods of prophylaxis, which include graduated compression stockings (GCS), and the use of intermittent pneumatic compression (IPC) devices and the venous foot pump (VFP), increase venous outflow and/or reduce stasis within the leg veins. The primary attraction of mechanical prophylaxis is the lack of bleeding potential. These modalities are, therefore, considerations for patients with high bleeding risks. While all three of the mechanical methods of prophylaxis have been shown to reduce the risk of DVT in a number of patient groups,^{2126127128129130131132133} they have been studied much less intensively than anticoagulant-based options and are generally less efficacious than the latter for the prevention of DVT.^{2131134135136}

No mechanical prophylaxis option has been shown to reduce the risk of death or PE. Special caution also should be exercised when interpreting the risk reductions ascribed to mechanical methods of prophylaxis for three reasons. Most trials were not blinded, increasing the chance of diagnostic suspicion bias. In the studies that used fibrinogen leg scanning to screen for DVT, mechanical prophylaxis may have factitiously lowered the 10 to 30% false-positive rate seen with the use of FUT (caused by venous pooling), while the rate remained unchanged in the nonmechanical treatment/control group.¹²⁶¹³⁷ Finally, because of relatively poor compliance with all mechanical options, they may not perform as well in routine clinical practice as in research studies in which major efforts are made to optimize proper use.^138139140 GCS should be used with caution in patients with arterial insufficiency.^141142143

In the recommendations that follow, the use of mechanical prophylaxis is an acceptable option in certain patient groups, especially in those patients who are at high risk for bleeding, or when used in combination with anticoagulant prophylaxis to improve efficacy.^133144145146 For all situations, the clinical staff must select the correct size of the device, must properly apply them,¹⁴⁷ and must ensure that they are removed for only a short time each day. Furthermore, nursing and physiotherapy initiatives should ensure that the devices do not impede ambulation.

Recommendation: Mechanical Methods of Prophylaxis
1.4.3. We recommend that mechanical methods of prophylaxis be used primarily in patients who are at high risk of bleeding (Grade 1C+), or as an adjunct to anticoagulant-based prophylaxis (Grade 2A). We recommend that careful attention be directed toward ensuring the proper use of, and optimal compliance with, the mechanical device (Grade 1C+).

1.4.4 Aspirin as thromboprophylaxis
Aspirin and other antiplatelet drugs are highly effective at reducing major vascular events in patients who are at risk for or who have established atherosclerotic disease.¹⁴⁸ Evidence^3149150151 suggests that antiplatelet agents also provide some protection against VTE in hospitalized patients who are at risk. However, we do not recommend the use of aspirin alone as VTE prophylaxis for several reasons. First, much of the evidence citing a benefit for the use of antiplatelet drugs against VTE is based on methodologically limited studies. For example, the Antiplatelet Trialists’ Collaboration metaanalysis¹⁴⁹ pooled data from generally small studies that were conducted > 25 years ago and that were of variable quality. Only one third of the studies included a group that received aspirin alone, and, of these, generally acceptable methods of screening for DVT were performed in only 38%.¹⁴⁹¹⁵² Second, a number of trials found no significant benefit from aspirin therapy,^{151153154155156} or found that aspirin was inferior to other prophylactic modalities.^2156157158 Finally, aspirin use is associated with a small but significant increased risk of major bleeding, especially if combined with other antithrombotic agents.¹⁴⁹¹⁵¹

The inferior efficacy of aspirin compared to other methods of VTE prophylaxis has been demonstrated in clinical trials. Among 205 patients undergoing hip or knee arthroplasty, who were randomized to receive aspirin or the LMWH ardeparin, the relative reduction in the risk of VTE with the use of LMWH over aspirin was 63% (p < 0.001).¹⁵⁷ The RRRs for DVT and proximal DVT in patients who have received prophylaxis with a VFP plus aspirin over that with aspirin alone following total knee arthroplasty (TKA) were 32% and > 95%, respectively (p < 0.001 for both comparisons).¹⁵⁶ Among hip fracture surgery (HFS) patients who were randomized to receive either aspirin or danaparoid, a low-molecular-weight heparinoid, VTE was detected in 44% and 28% of the patients, respectively (p = 0.028).¹⁵⁸

Recommendation: Aspirin
1.4.4. We recommend against the use of aspirin alone as prophylaxis against VTE for any patient group (Grade 1A).

1.4.5 Application of evidence to individual patients
The prophylaxis recommendations contained in this report apply to groups of patients for whom the benefits of prophylaxis appear to outweigh the risks. Decisions about prescribing prophylaxis for the individual patient are best made by combining knowledge of the literature (including the recommendations provided herein) with clinical judgment, the latter based on specific knowledge about each patient’s risk factors for VTE, the potential for adverse consequences with prophylaxis, and the availability of various options within one’s center. Since most thromboprophylaxis studies excluded patients who were at high risk for either VTE or adverse outcomes, their results may not apply to those patients with previous VTE or who have an increased risk of bleeding. In these circumstances, clinical judgment may appropriately warrant the use of a prophylaxis option that differs from the recommended approach.

Renal clearance is the primary mode of elimination for several anticoagulants, including LMWH, fondaparinux, and the direct thrombin inhibitor melagatran. With reduced creatinine clearance, these drugs may accumulate and increase the risk of bleeding.¹⁵⁹¹⁶⁰ However, each agent must be evaluated separately since there appears to be considerable variability in the relationship between renal impairment and drug accumulation even for various LMWHs.¹⁶¹

Recommendations: Dosing and Renal Impairment
1.4.5.1. For each of the antithrombotic agents, we recommend that clinicians consider the manufacturer’s suggested dosing guidelines (Grade 1C).

1.4.5.2. We recommend consideration of renal impairment when deciding on doses of LMWH, fondaparinux, the direct thrombin inhibitors, and other antithrombotic drugs that are cleared by the kidneys, particularly in elderly patients and those who are at high risk for bleeding (Grade 1C+).

1.5 Antithrombotic drugs and neuraxial anesthesia/analgesia
The benefits of neuraxial blockade (ie, spinal or epidural anesthesia and continuous epidural analgesia) are well-established.^{162163164165166167} The risk of perispinal hematoma, a very rare but potentially devastating complication after neuraxial blockade, may be increased with the concomitant use of antithrombotic drugs.¹⁶⁸¹⁶⁹ Bleeding into the enclosed space of the spinal canal can produce spinal cord ischemia and subsequent paraplegia. The seriousness of this complication mandates the cautious use of all antithrombotic medications in patients undergoing neuraxial blockade. A 1997 Food and Drug Administration public health advisory¹⁷⁰¹⁷¹ reported 41 US patients who developed perispinal hematoma after receiving the LMWH enoxaparin around the time of spinal/epidural anesthesia. Some patients had preexisting spinal abnormalities, and a third had received additional hemostasis-inhibiting medications. Nearly 90% of the cases occurred among patients receiving enoxaparin as thromboprophylaxis after knee or hip replacement or after spinal surgery. Many of these patients experienced neurologic impairment, including permanent paralysis, despite undergoing a decompressive laminectomy. Additional cases of perispinal hematoma in patients who have received LMWH have been reported. This complication also has been reported with the use of LDUH, although apparently with lower frequency.

Most patients who develop perispinal hematomas have more than one risk factor for local or systemic bleeding, including the presence of an underlying hemostatic disorder, anatomic or vascular vertebral column abnormalities, traumatic needle or catheter insertion, repeated insertion attempts, insertion in the presence of high levels of an anticoagulant, the use of continuous epidural catheters, the concurrent administration of medications known to increase bleeding, high anticoagulant dosage, older age, and female gender.^168170171 Removal of the epidural catheter, especially in the presence of an anticoagulant effect, has also been associated with hematoma.¹⁶⁸ Unfortunately, the prevalence of perispinal hematoma and the predictive value of the various risk factors remain unknown. As a result, reviews on the use of antithrombotic therapy among recipients of neuraxial anesthesia^169172173 combine the limited available evidence with practical advice. A detailed discussion of this topic is available through the American Society of Regional Anesthesia and Pain Medicine (www.asra.com).¹⁶⁹

Consideration of neuraxial anesthesia plus or minus postoperative epidural analgesia requires a review of the intended benefits and the potential risks. A careful history will identify most patients with an important underlying bleeding disorder and those receiving agents that affect hemostasis or platelet function. In keeping with the American Society of Regional Anesthesia recommendations, we believe that neuraxial blockade and anticoagulant thromboprophylaxis, including the use of LDUH and LMWH, can generally be used concurrently as long as there is appropriate caution.

The following suggestions may improve the safety of neuraxial blockade in patients who have or will receive anticoagulant prophylaxis. (1) Neuraxial anesthesia/analgesia should generally be avoided in patients with a known bleeding disorder. (2) Neuraxial anesthesia should generally be avoided in patients whose preoperative hemostasis is impaired by antithrombotic drugs. Nonsteroidal anti-inflammatory agents and aspirin do not appear to increase the risk of perispinal hematoma. Since less is known about the safety of the thienopyridine platelet inhibitors clopidogrel and ticlopidine in patients undergoing neuraxial block, the discontinuation of these drugs 5 to 14 days before the procedure should be considered. In patients receiving preoperative anticoagulants, the insertion of the spinal needle or epidural catheter should be delayed until the anticoagulant effect of the medication is minimal. This is usually at least 8 to 12 h after a subcutaneous dose of heparin or a twice daily prophylactic dose of LMWH, or at least 18 h after a once-daily LMWH injection. (3) Anticoagulant prophylaxis should be delayed if a hemorrhagic aspirate (ie, a "bloody tap") is encountered during the initial spinal needle placement. (4) Removal of an epidural catheter should be done when the anticoagulant effect is at a minimum (usually just before the next scheduled subcutaneous injection). (5) Anticoagulant prophylaxis should be delayed for at least 2 h after spinal needle or epidural catheter removal. (6) If prophylaxis with a VKA, such as warfarin, is used, we recommend that continuous epidural analgesia not be used for longer than 1 or 2 days because of the unpredictable anticoagulant effect of the anticoagulant. Furthermore, if prophylaxis with a VKA is used at the same time as epidural analgesia, the international normalized ratio (INR) should be < 1.5 at the time of catheter removal. (7) Although postoperative prophylaxis with fondaparinux appears to be safe in patients who have received a spinal anesthetic, there are no safety data about its use along with postoperative continuous epidural analgesia. The long half-life of fondaparinux and its renal mode of excretion raise concerns about the potential for accumulation of the drug, especially in the elderly because of the associated impairment of renal function. Until further data are available, we recommend that fondaparinux not be administered along with continuous epidural analgesia.

With the concurrent use of epidural analgesia and anticoagulant prophylaxis, all patients should be monitored carefully and frequently for the symptoms and signs of cord compression. These symptoms include progression of lower extremity numbness or weakness, bowel or bladder dysfunction, and new onset of back pain. If spinal hematoma is suspected, diagnostic imaging and definitive surgical therapy must be performed rapidly to reduce the risk of permanent paresis. We encourage every hospital that uses neuraxial anesthesia/analgesia to develop written protocols that cover the most common scenarios in which these techniques will be used along with antithrombotic agents.

Recommendation: Neuraxial Anesthesia/analgesia
1.5.1. In all patients undergoing neuraxial anesthesia or analgesia, we recommend special caution when using anticoagulant prophylaxis (Grade 1C+).

2.0 General, Vascular, Gynecologic, and Urologic Surgery
2.1 General surgery
In studies published between 1969 and 1984,^4077174 the observed rate of DVT among general surgical patients not receiving prophylaxis varied between 15% and 30%, with rates of fatal PE between 0.2% and 0.9%.The current risk of thromboembolic complications in general surgery is unknown because studies without prophylaxis are no longer performed in these patients. More rapid mobilization, greater use of thromboprophylaxis, and other advances in perioperative care may tend toward reducing the thromboembolic risk. However, the performance of more extensive operative procedures in older and sicker patients, the use of preoperative chemotherapy, and the shorter lengths of stay in the hospital (leading to shorter durations of prophylaxis) may heighten the risk of VTE in contemporary patients undergoing inpatient, general surgery.

The type and duration of surgery clearly influence the risk of DVT.^{90175176177178} Most individuals undergoing outpatient surgery appear to have a low frequency of DVT. For example, only one case of symptomatic VTE arose in the first month following 2,281 day-case hernia repairs (0.04%).¹⁷⁹ Additional factors that alter the risk of VTE in general surgery patients include the following:

Traditional risk factors such as cancer, previous VTE, obesity, varicose veins, and estrogen use^175176177178;

Increasing age, an independent risk factor for VTE¹⁷⁶¹⁷⁷;

Type of anesthesia. In the absence of prophylaxis, the risk of DVT is lower following spinal/epidural anesthesia than after general anesthesia.¹⁸⁰ This protective effect is less apparent, at least in orthopedic surgery, when pharmacologic prophylaxis is used¹⁸¹¹⁸²; and

General perioperative care, including degree of mobilization, fluid status, and transfusion practices.

Furthermore, the diagnostic screening test used (ie, FUT, venography, or DUS) and the quality of its interpretation greatly affect the rate of detection of thrombi, as discussed in section 1.4.1.^{9599100103110123}

Based on the results of numerous randomized clinical trials and metaanalyses, we recommend the routine use of thromboprophylaxis following major general surgical procedures.¹²³⁸⁹ Both LDUH and LMWH reduce the risk of both asymptomatic and symptomatic VTE in general surgery by at least 60%.²⁷¹⁷⁷ Most prophylaxis trials of subcutaneous LDUH administered 5,000 U 1 to 2 h before surgery, followed by administration of 5,000 U bid or tid until patients were either ambulating or were discharged from hospital. A metaanalysis of 46 randomized clinical trials in general surgery⁷¹ compared therapy with LDUH with no prophylaxis or placebo. The rate of DVT was significantly reduced (from 22 to 9%; odds ratio [OR], 0.3; NNT, 7), as were the rates of symptomatic PE (from 2.0 to 1.3%; OR, 0.5; NNT, 143), fatal PE (from 0.8 to 0.3%; OR, 0.4; NNT, 182), and all-cause mortality (from 4.2 to 3.2%; OR, 0.8; NNT, 97). Prophylaxis with LDUH was associated with a small increase in the rate of bleeding events (from 3.8 to 5.9%; OR, 1.6; NNH, 47). These findings were verified in another metaanalysis⁷⁷ in which the rate of wound hematomas was increased with LDUH use (6.3% vs 4.1% in control subjects; OR, 1.6; NNH, 45), although the rate of major bleeding was not. While both meta-analyses concluded that the administration of heparin, 5,000 U tid, was more efficacious than that of 5,000 U bid, without increasing bleeding, this was based on indirect comparisons, and we are not aware of any studies that directly compared these two regimens.

LMWHs have been evaluated extensively in general surgery patients, usually in comparison with LDUH.^{79102183184185186187188189190191192193194195196197198199} Clinical trials also have compared different LMWHs²⁰⁰ or different regimens of the same LMWH.^{101188190201202203204205206} One metaanalysis⁴⁰ found that LMWH prophylaxis reduced the risk of asymptomatic DVT and symptomatic VTE in general surgery patients by > 70% compared with no prophylaxis.

When LDUH and LMWH were directly compared, no single study showed a difference in the rates of symptomatic VTE, although several trials^183185187 found that LMWH was associated with significantly fewer asymptomatic DVTs. There are at least nine metaanalyses and systematic reviews comparing these two agents.^{4078808182207208209210} Small differences in their results can be explained by the variability in the inclusion criteria for the original studies. The various LMWHs were grouped together as a single class agent, despite differences in their pharmacologic properties and the position statements made by regulatory authorities that each LMWH should be considered as a distinct drug. We are not aware of any direct comparisons of comparable doses of different LMWHs in this patient population.²⁰⁰

In summary, for general surgery patients, LDUH and LMWHs have similar efficacy and bleeding rates. In high-risk general surgery patients, higher doses of LMWH provide greater protection than lower doses.^101195211212 For example, in cancer patients, prophylaxis with dalteparin, 5,000 U daily, was significantly more efficacious than with 2,500 U daily, without an increased risk of bleeding.¹⁰¹

Some studies^79102193196 have reported significantly fewer wound hematomas and other bleeding complications with LMWH than with LDUH, while other trials^184185199 have shown the opposite effect. Two meta-analyses⁴⁰⁸¹ that found similar efficacy for LDUH and LMWH described differences in bleeding rates that were dependent on the dose of LMWH used. Lower doses of LMWH (ie, 3,400 U daily) were associated with less bleeding than LDUH (3.8% vs 5.4%, respectively; OR, 0.7), while higher doses resulted in more bleeding events (7.9% vs 5.3%, respectively; OR, 1.5).⁸¹

The clinical advantages of LMWH over LDUH include its once-daily administration and the lower risk of heparin-induced thrombocytopenia (HIT),²¹³ while, at least in North America, LMWH is more costly.

Several large studies in general surgery patients have evaluated the risk of death among patients given LDUH or LMWH. Two clinical trials⁵⁰⁶⁹ were specifically designed to test the effectiveness of LDUH in preventing fatal PE, compared with no prophylaxis. Both studies demonstrated a significant beneficial effect (overall RRR for fatal PE with LDUH, 91%; NNT, 106).⁵⁰⁶⁹ A placebo-controlled, multicenter study¹⁷⁴ found that the LMWH fraxiparine significantly reduced the all-cause mortality rate (from 0.8 to 0.4%) among 4,498 general surgery patients (NNT, 250). Two additional randomized clinical trials,¹⁹¹¹⁹⁷ with a combined sample of 35,000 surgical patients, found no difference in the rates of total mortality, fatal PE, or bleeding between LDUH (5,000 U tid) and the LMWH certoparin (3,000 U once daily). In both studies, the follow-up duration was brief (14 days and the in-hospital period only).

The selective inhibitor of factor Xa fondaparinux has been evaluated in a randomized, double-blinded clinical trial²¹⁴ among almost 3,000 patients undergoing high-risk abdominal surgery. Prophylaxis with fondaparinux, started postoperatively, was compared with prophylaxis with dalteparin started before surgery. There were no significant differences in the rates of VTE, major bleeding, or death between the two prophylaxis groups.

Although mechanical methods of prophylaxis (ie, GCS and IPC) are attractive options in general surgery patients who have a high risk of bleeding, they have not been studied as extensively as has pharmacologic prophylaxis.⁷⁷ A systematic review¹³³ observed a significant 52% reduction in the rate of DVT with the use of GCS (13%) compared with no prophylaxis (27%), which is equivalent to a pooled OR of 0.3 (NNT, 7). This finding was confirmed by two additional meta-analyses.¹³⁰²¹⁵ The use of GCS has also been shown to enhance the protective effect of LDUH against DVT by a further 75% compared with LDUH alone (DVT rates of 4% and 15% in the combined and LDUH groups, respectively), for a pooled OR of 0.2 (NNT, 9).¹³³ The effect of GCS on the risk of proximal DVT or symptomatic PE, and their effectiveness in patients with malignancies remains unknown due to the presence of only a few small studies. Some practical limitations of GCS include a lack of standardization of the quality of the stockings, difficulty with fitting patients with unusual limb sizes or shapes, and poor compliance with their use by both health-care providers and patients.¹³⁸¹⁴⁰

Several small, older studies^216217218 have suggested that prophylaxis with IPC might reduce the incidence of DVT in general surgical patients to an extent similar to LDUH, although another study²¹⁹ found that IPC provided no protection at all. There is insufficient evidence to assess whether IPC prophylaxis alone has any effect on symptomatic VTE or mortality. In a single randomized clinical trial of 2,551 cardiac surgery patients,¹⁴⁶ the rate of symptomatic PE was lower with combined IPC and LDUH (1.5%) than with LDUH alone (4.0%).

Although the risk of developing postoperative DVT is highest within the first week or two after undergoing general surgery, VTE complications, including fatal PE, may occur later.^914177220 In one study,²²¹ 51 patients who underwent major abdominal surgery received thromboprophylaxis in the hospital and had DVT excluded at the time of hospital discharge. Follow-up at home over the next 4 weeks, using serial FUT and DUS, detected DVT in 13 patients (25%). These studies and the current brief lengths of hospital stay have prompted assessments of the optimal duration of prophylaxis following general surgical procedures.

Three clinical trials^204205206 have addressed the use of extended prophylaxis beyond the period of hospitalization following general surgery. In a small, nonblinded trial in 118 major abdominal or thoracic surgery patients, 4 weeks of tinzaparin, 3,500 U daily, was associated with a nonsignificant reduction in the risk of asymptomatic DVT detected by bilateral screening venography, compared with the same dose given for just 1 week (DVT rates, 5% and 10%, respectively).²⁰⁴ In another open-label study conducted in 233 major abdominal surgery patients,²⁰⁶ dalteparin, 5,000 U, was administered once daily for 1 or 4 weeks. Bilateral venography detected DVT in 16% and 6%, respectively, of the patients who received prophylaxis for 1 week or 1 month (p = 0.09) [proximal DVT rates, 9% and 0%, respectively; p = 0.001]. A subgroup analysis²²² of the patients in this study who had malignancies reported statistically significant RRRs in the rates of DVT and proximal DVT with extended prophylaxis. The ENOXACAN II study,²⁰⁵ a double-blinded, multicenter trial conducted in 332 abdominal or pelvic cancer surgery patients, compared the administration of enoxaparin, 40 mg daily, for an average of 9 or 28 days. Routine venography, performed between days 25 and 31, showed a significant reduction in DVT rates with the prolonged prophylaxis (from 12 to 5%; OR, 0.36; p = 0.02). However, proximal DVT was identified in only three patients in the short-duration group and in one patient in the extended prophylaxis group. Over the entire 3-month follow-up period, there were only two symptomatic thromboembolic events among the short-duration patients and one event in the extended prophylaxis group.

In conclusion, among patients undergoing major general surgical procedures, routine thromboprophylaxis is recommended.^1238789 The options that have clearly been shown to reduce DVT and PE are LDUH and LMWH. Mechanical prophylactic methods (ie, GCS and/or IPC) appear to reduce DVT rates and should be considered for patients who are at a particularly high risk of bleeding. Prophylaxis with LMWH for 2 to 3 weeks after hospital discharge appears to reduce the incidence of asymptomatic DVT in cancer surgery patients.

Recommendations: General Surgery
2.1.1. In low-risk general surgery patients (Table 5) who are undergoing a minor procedure, are < 40 years of age, and have no additional risk factors, we recommend against the use of specific prophylaxis other than early and persistent mobilization (Grade 1C+).

2.1.2. Moderate-risk general surgery patients are those patients undergoing a nonmajor procedure and are between the ages of 40 and 60 years or have additional risk factors, or those patients who are undergoing major operations and are < 40 years of age with no additional risk factors. We recommend prophylaxis with LDUH, 5,000 U bid or LMWH 3,400 U once daily (both Grade 1A).

2.1.3. Higher-risk general surgery patients are those undergoing nonmajor surgery and are > 60 years of age or have additional risk factors, or patients undergoing major surgery who are > 40 years of age or have additional risk factors. We recommend thromboprophylaxis with LDUH, 5,000 U tid or LMWH, > 3,400 U daily (both Grade 1A).

2.1.4. In high-risk general surgery patients with multiple risk factors, we recommend that pharmacologic methods (ie, LDUH, tid or LMWH, > 3,400 U daily) be combined with the use of GCS and/or IPC (Grade 1C+).

2.1.5. In general surgery patients with a high risk of bleeding, we recommend the use of mechanical prophylaxis with properly fitted GCS or IPC, at least initially until the bleeding risk decreases (Grade 1A).

2.1.6. In selected high-risk general surgery patients, including those who have undergone major cancer surgery, we suggest post-hospital discharge prophylaxis with LMWH (Grade 2A).

2.2 Vascular surgery
Most patients undergoing vascular surgery routinely receive one or more antithrombotic agents to prevent vascular occlusion. This is achieved using perioperative platelet inhibitors, such as aspirin or clopidogrel, and intraoperative heparins or dextran before vascular clamping. Postoperative anticoagulation therapy with unfractionated heparin, warfarin, and/or LMWH is also common in these patients.^223224225 Because of the widespread use of these agents, little is known about the frequency of VTE in vascular surgery patients, especially among those not receiving antithrombotic drugs. In a population-based study²⁰ of 1.6 million patients, the incidence of symptomatic VTE within 3 months of major vascular surgery was 1.7 to 2.8%. Potential thromboembolic risk factors in vascular surgery include advanced age, limb ischemia, long duration of surgery, and intraoperative local trauma, including possible venous injury.⁶ Preliminary evidence²²⁶ suggests that atherosclerosis also may be an independent risk factor for VTE.

The 20 to 30% rate of asymptomatic DVT after aortoiliac or aortofemoral surgery is similar to that reported in other abdominal and pelvic procedures.^227228229230 However, these rates may have been inflated by the high false-positive rates (25 to 81%) seen with FUT, which were clearly identified when patients with abnormal FUT results also underwent venography.^228231232 In five prospective studies of vascular surgery patients not receiving any thromboprophylaxis, the pooled rate of postoperative DVT was 21% (18 of 86 patients) using contrast venography^233234235 and 15% (15 of 98 patients) using DUS.²³⁰²³⁶ In another study of 50 patients undergoing aortic aneurysm repair,²³⁵ asymptomatic DVT was diagnosed in 18% of patients using contrast venography, while the rate of proximal DVT was 4%. Among 142 patients who underwent a variety of vascular surgical procedures, all of whom received thromboprophylaxis with intraoperative IPC and perioperative LDUH, the respective rates of DVT and proximal DVT, which were detected by routine screening with DUS on days 7 to 10, were 10% and 6%, respectively.²³⁷

Aortic aneurysm resection or aortofemoral bypass appears to confer a higher risk of DVT than femorodistal bypass. We are aware of only three studies that routinely screened for DVT and also included both groups of patients.^230237238 In one randomized trial,²³⁸ patients received either subcutaneous LDUH or LMWH. Using DUS screening at days 7 to 10 after surgery, with venography confirmation of a positive DUS result, DVT was detected in 8% of patients (11 of 146 patients) who underwent aortic surgery and in 3% of those who underwent femorodistal bypass (3 of 87 patients). In a second study,²³⁷ routine DUS was performed in vascular surgery patients, who also received prophylaxis with IPC and LDUH. The respective rates of DVT were 12% (6 of 52 patients) and 9% (5 of 54 patients), respectively, among the patients who had aortic and femorodistal surgery. In the most recent study,²³⁰ a pre-hospital discharge DUS was obtained in 50 vascular surgery patients, none of whom had received thromboprophylaxis. Again, the rate of DVT was higher in the aortic surgery patients (41%) than in the peripheral arterial surgery patients (18%). A prospective registry²³⁹ of 7,533 vascular surgery procedures performed in Finland reported clinical DVT in 0.9% of patients after aortic surgery and 0.7% after femorodistal reconstruction.

In patients with lower limb ischemia, preoperative DVT may be present. One study detected DVT by DUS in 20% of 136 peripheral vascular disease patients prior to arteriography or surgery, although no DVT appeared to be acute by ultrasonographic assessment.²⁴⁰ Logistic regression analysis showed that increased severity of ischemia, expressed as a reduced ankle pressure/brachial pressure index, was the only independent risk factor for DVT. In another prospective study,²³⁰ only 1 of 53 vascular surgery patients was found to have DVT on preoperative DUS. A third DUS-based study²⁴¹ reported low rates of preoperative asymptomatic DVT (4%) and postoperative asymptomatic DVT (3%) in patients undergoing infrainguinal arterial reconstruction, although 25% of the patients received anticoagulation therapy postoperatively. Even patients who have had endovascular treatment of abdominal aortic aneurysms are at risk for DVT. For example, 6% of 50 consecutive patients who had DUS on the first and 30th days postprocedure were found to have DVTs.²⁴²

Only four randomized clinical trials of thromboprophylaxis after arterial reconstructive surgery have been performed.^228236238243 In each of the studies, patients received IV heparin during the procedure. The first trial²²⁸ compared LDUH, 5,000 U bid, to placebo in 49 patients undergoing elective aortic bifurcation surgery. Using FUT, confirmed by venography if positive, DVT was detected in 24% of placebo recipients and in 4% of LDUH recipients. However, clinical bleeding was significantly greater in those who received LDUH, leading to the premature termination of the study. A second study²⁴³ found no benefit of LDUH over no prophylaxis, although only 43 patients were included. In the third trial,²³⁶ 100 patients having aortic surgery were randomized to receive LDUH plus GCS or no prophylaxis. Proximal DVT was detected in 2% of patients in both groups using serial DUS. The final study²³⁸ compared LDUH, 7,500 U bid, with enoxaparin, 40 mg daily, with each administered for 2 days, among 233 patients undergoing aortic or infrainguinal reconstructions. Using DUS at days 7 to 10, DVT was detected in 4% and 8% of patients, respectively (not statistically significant), while major bleeding occurred in 2% of patients in both groups.

Recommendations: Vascular Surgery
2.2.1. In patients undergoing vascular surgery who do not have additional thromboemobolic risk factors, we suggest that clinicians not routinely use thromboprophylaxis (Grade 2B).

2.2.2. For patients undergoing major vascular surgical procedures who have additional thromboembolic risk factors, we recommend prophylaxis with LDUH or LMWH (Grade 1C+).

2.3 Gynecologic surgery
VTE is an important and potentially preventable complication of major gynecologic surgery, with rates of DVT, PE, and fatal PE comparable to those seen after general surgical procedures.^2244245 Several factors appear to increase the risk of VTE following gynecologic surgery, including malignancy, older age, previous VTE, prior pelvic radiation therapy, and use of an abdominal surgical approach.^20246247 Gynecologic oncology patients are often elderly, and they all have cancer, with or without compression of the major pelvic veins by a mass.²⁴⁶²⁴⁸ Venous intimal injury may occur following preoperative radiotherapy or during surgery (especially with pelvic lymph node dissection), the procedures are frequently lengthy, and residual tumor may be left in situ. Postoperative mobility is often impaired after such extensive surgery, and chemotherapy itself is thrombogenic. As in other surgical patients, thrombi generally form during or shortly after the procedure,²⁴⁹ although most symptomatic thromboemboli occur after hospital discharge.²⁴⁷²⁵⁰

Despite substantial changes in surgical and postoperative care, few randomized clinical trials of thromboprophylaxis in gynecologic surgery have been reported in the past decade,^{245251252253254255} and some of these have major methodological limitations.

Several practice guidelines have addressed the issue of thromboprophylaxis in patients undergoing gynecologic surgery.^2256257 Patients who are otherwise well and undergo brief procedures, typically defined as < 30 min, do not require any specific prophylaxis but should be encouraged to mobilize early after surgery. The previous American College of Chest Physicians Consensus Conference on Antithrombotic Therapy concluded that twice daily dosing of LDUH was effective in patients undergoing gynecologic surgery for benign disease in the absence of additional risk factors.² Mechanical prophylaxis with IPC also appears to be efficacious in this group^251258259 and should be considered for patients who are at a high risk of bleeding. IPC prophylaxis should be started just before surgery, used continuously while the patient is not ambulating, and stopped just before hospital discharge. Formal strategies to optimize compliance with IPC by patients and nursing staff are essential.

Patients having surgery for gynecologic cancers appear to derive less protection from twice daily dosing of LDUH than those with benign disease.²⁶⁰²⁶¹ LDUH, given three times daily, or LMWH, at daily doses of at least 4,000 U, appear to be more effective in these cancer patients.^{195251255261262} Four randomized clinical trials^195248254263 compared LDUH, given three times daily, with LMWH in gynecologic surgery patients, and suggested similar effectiveness and safety with either approach. In an uncontrolled case series of 2,030 patients who were undergoing major gynecologic surgery and were given enoxaparin, 20 mg once daily, there were no fatal PEs, and only seven patients (0.3%) developed symptomatic VTE.²⁶⁴ Combining mechanical prophylaxis with LDUH or LMWH therapy may enhance efficacy, although, to our knowledge, this has not been studied in gynecology patients.

Although the risk of VTE after laparoscopic gynecologic surgery is unknown (and appears to be lower than that for open procedures), laparoscopic procedures result in impaired venous return from the legs and activation of coagulation. Therefore, we recommend that a decision to provide prophylaxis be individualized, considering a patient’s comorbid and procedure-related risk factors.

Another unresolved issue is the duration of antithrombotic prophylaxis following gynecologic surgery. One randomized, double-blind study²⁰⁵ compared 1 week with 1 month of LMWH prophylaxis in patients undergoing curative surgery for abdominal or pelvic malignancy (8% of the patients had a gynecologic oncology procedure). Extended prophylaxis conferred a RRR of 60% for both venographically screened DVT and proximal DVT. While this trial suggested a potential advantage of post-hospital discharge prophylaxis in certain high-risk surgical oncology patients, the specific risk factors that warrant extended prophylaxis remain to be defined. In a recent study of 1,862 patients who underwent gynecologic surgery and received IPC prophylaxis,²⁴⁷ the risk factors for symptomatic VTE included cancer surgery, previous DVT, and age > 60 years.

Recommendations: Gynecologic Surgery
2.3.1. For gynecologic surgery patients undergoing brief procedures of 30 min for benign disease, we recommend against the use of specific prophylaxis other than early and persistent mobilization (Grade 1C+).

2.3.2. For patients undergoing laparoscopic gynecologic procedures, in whom additional VTE risk factors are present, we recommend the use of thromboprophylaxis with one or more of the following: LDUH, LMWH, IPC, or GCS (all Grade 1C).

2.3.3. We recommend that thromboprophylaxis be used in all major gynecologic surgery patients (Grade 1A).

2.3.4. For patients undergoing major gynecologic surgery for benign disease, without additional risk factors, we recommend LDUH, 5,000 U bid (Grade 1A). Alternatives include once-daily prophylaxis with LMWH, 3,400 U/d (Grade 1C+), or IPC started just before surgery and used continuously while the patient is not ambulating. (Grade 1B).

2.3.5. For patients undergoing extensive surgery for malignancy, and for patients with additional VTE risk factors, we recommend routine prophylaxis with LDUH, 5,000 U tid (Grade 1A), or higher doses of LMWH (ie, > 3,400 U/d) [Grade 1A]. Alternative considerations include IPC alone continued until hospital discharge (Grade 1A), or a combination of LDUH or LMWH plus mechanical prophylaxis with GCS or IPC (all Grade 1C).

2.3.6. For patients undergoing major gynecologic procedures, we suggest that prophylaxis continue until discharge from the hospital (Grade 1C). For patients who are at particularly high risk, including those who have undergone cancer surgery and are > 60 years of age or have previously experienced VTE, we suggest continuing prophylaxis for 2 to 4 weeks after hospital discharge (Grade 2C).

2.4 Urologic surgery
VTE is considered to be the most important nonsurgical complication following major urologic procedures.^{2265266267268269} Unfortunately, most of the epidemiologic data related to VTE in this population were derived 10 to 30 years ago. Subsequent changes in surgical care, earlier mobilization, and possibly greater use of prophylaxis have been associated with declining rates of VTE over time.^270271272 However, 1 to 5% of contemporary patients undergoing major urologic surgery experience symptomatic VTE, with PE believed to be the most common cause of postoperative death, at a risk of < 1 in 500.^{20266270271272273274275276277278279280}

Patients undergoing major urologic surgery often have multiple risk factors for VTE, including advanced age, malignancy, use of the lithotomy position intraoperatively, and pelvic surgery with or without lymph node dissection. Additional factors for DVT include the use of open (vs transurethral) procedures and a longer duration of the procedure.

Most of the information about VTE and its prevention were derived from patients undergoing open prostatectomy. Other urologic procedures, including major renal surgery and transplantation, radical cystectomy, and urethral reconstruction, are also associated with an increased risk for thrombosis and warrant consideration for prophylaxis.^281282283

We identified only one randomized clinical trial of thromboprophylaxis in urologic surgery published over the past 2 decades that met the minimal methodological criteria (Table 1).²⁸⁴ Thus, the optimal approach to thromboprophylaxis in these patients is not known.²²⁸⁵ Many older studies were small, and lacked blinding and objective outcome assessments. Modern anesthesia techniques have improved, and there is generally a more aggressive approach to postoperative mobilization. At the same time, radical cancer operations are being performed more frequently than in the past. Despite a sparse literature on thromboprophylaxis in patients undergoing urologic surgery, the risks of VTE and the protection offered by various prophylaxis methods appear to be similar to those seen in patients undergoing major general or gynecologic surgery.²³⁷¹ Furthermore, consideration of bleeding risk is particularly important in urologic surgery, especially following prostatectomy.

While the use of GCS or IPC prophylaxis is likely to be efficacious in urologic surgery,^{126274280284286} IPC is more expensive and may provide no additional protection over the use of GCS alone.²⁷⁴²⁸⁶ Both LDUH and LMWH are efficacious in patients undergoing urologic surgery.^{71264268279287288289} Concerns about the potential for pelvic hematomas and lymphoceles in patients receiving anticoagulant prophylaxis have been raised by some investigators,^268272290 but not by others.^268279289 The combination of mechanical and pharmacologic prophylaxis may be more effective than either alone but may not be necessary and is more expensive.^268279291

For patients undergoing transurethral prostatectomy, the risks of VTE are low,^{2071264280289292} and perioperative use of LDUH or LMWH may increase the risk of bleeding.^290293294 Early postoperative mobilization is probably the only intervention warranted in these and other low-risk urologic surgery patients. Routine prophylaxis is recommended for more extensive, open procedures including radical prostatectomy, cystectomy, or nephrectomy. Until further data are available, VTE prophylaxis options to consider for these patients include the following: LDUH; LMWH; GCS; and IPC. For urology patients who are at particularly high risk, commencing prophylaxis with GCS with or without IPC just prior to surgery and then adding LDUH or LMWH postoperatively should be considered, even though this approach has not been formally evaluated in this patient population. With the current brief lengths of hospitalization, even for major urologic procedures, the risk of post-hospital discharge, symptomatic VTE is likely to increase.²⁰²⁹⁵ Therefore, the optimal duration of prophylaxis is uncertain. Patients who are believed to be at high risk for thromboembolism, including elderly patients undergoing radical prostatectomy, patients with a history of VTE, or patients who have limited mobility at hospital discharge, should be considered for post-hospital discharge thromboprophylaxis.²⁰⁵

Recommendations: Urologic Surgery
2.4.1. In patients undergoing transurethral or other low-risk urologic procedures, we recommend against the use of specific prophylaxis other than early and persistent mobilization (Grade 1C+).

2.4.2. For patients undergoing major, open urologic procedures, we recommend routine prophylaxis with LDUH twice daily or three times daily (Grade 1A). Acceptable alternatives include prophylaxis with IPC and/or GCS (Grade 1B) or LMWH (Grade 1C+).

2.4.3. For urologic surgery patients who are actively bleeding, or are at very high risk for bleeding, we recommend the use of mechanical prophylaxis with GCS and/or IPC at least until the bleeding risk decreases (Grade 1C+).

2.4.4. For patients with multiple risk factors, we recommend combining GCS and/or IPC with LDUH or LMWH (Grade 1C+).

2.5 Laparoscopic surgery
The expanding use of laparoscopic techniques over the past 2 decades has profoundly changed surgical diagnosis and therapy. There is, however, considerable controversy related to thromboembolic complications after these procedures.²⁹⁶ Laparoscopic cholecystectomy is associated with a modest thrombogenic activation of the coagulation system,^{297298299300301302303304} as well as stimulation of fibrinolysis.³⁰⁵³⁰⁶ In some studies,²⁹⁸³⁰⁵ the magnitude of these changes was similar to that of changes seen after open cholecystectomy, while other studies^299302306 found smaller changes among the patients undergoing laparoscopic cholecystectomy. Laparoscopic operations are often associated with longer surgical times than are open procedures. Both pneumoperitoneum and the reverse Trendelenburg position reduce venous return from the legs, creating lower extremity venous stasis.^307308309 While laparoscopic procedures are generally associated with a shorter hospital stay, patients undergoing them may not mobilize more rapidly at home than those undergoing open procedures.

Although the risks of VTE and its prevention have been less intensively studied in laparoscopic procedures compared with other abdominal procedures, the risks appear to be low.²⁰ For example, among 417 UK surgeons, 91% reported having never encountered a thromboembolic complication following laparoscopic cholecystectomy, although the majority reported using LDUH routinely in these patients.³¹⁰ A Danish survey³¹¹ found that 80% of surgical departments were not aware of any thromboembolic complications following laparoscopic surgery, although, again, prophylaxis was commonly used. In another study, no DVT or PE was encountered in the first month after laparoscopic cholecystectomy among 587 cases, of whom only 3% received thromboprophylaxis.³¹²

Among 25 patients undergoing laparoscopic cholecystectomy without any thromboprophylaxis, screening contrast venography on postoperative days 6 to 10, failed to detect any DVT.³¹³ Eight cases of DVT (0.3%) and no cases of PE were seen in another series of 2,384 consecutive patients who underwent GI laparoscopic procedures followed by a short course of LMWH prophylaxis.³¹⁴ A review of 50,427 gynecologic laparoscopies³¹⁵ observed a symptomatic VTE rate of only 2 per 10,000 patients. In a literature review of laparoscopic cholecystectomy³¹⁶ including 11,863 patients, only 3 of the 10 postoperative deaths were attributed to PE. In another literature review of 153,832 laparoscopic cholecystectomies,³¹⁷ using various types of prophylaxis, the average rates of clinical DVT, PE, and fatal PE were 0.03%, 0.06%, and 0.02%, respectively. In a prospective national Swedish registry,³¹⁸ VTE was encountered in only 0.2% of the 11,164 patients who underwent laparoscopic cholecystectomies. However, the proportion of patients who received thromboprophylaxis was not reported. Finally, in a population-based study of 105,850 laparoscopic cholecystectomies performed in California,²⁰ the risk of symptomatic VTE within 3 months of the procedure was 0.2%, compared with 0.5% after open cholecystectomy.

Table 6 shows the rates of objectively proven DVT after laparoscopy, which were derived from prospective studies that used various forms of prophylaxis.^{297313319320321322323324325} Although the studies were generally small, with a single exception the rates of asymptomatic DVT were very low. Among the eight prospective studies that used routine postoperative DUS, the pooled rate of DVT was 1.4% (17 of 1,248 patients). Excluding one outlier study, the DVT rate was 0.5% among the 1,228 patients. When no prophylaxis was given, the rate of asymptomatic DVT in the 219 patients rose to 0.9%.