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 Full Text (PDF) Free Submit a response View responses 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 Wheeler, H. B. Search for Related Content PubMed PubMed Citation Articles by Geerts, W. H. Articles by Wheeler, H. B.

Prevention of Venous Thromboembolism

William H. Geerts, MD, FCCP, Chair^*; John A. Heit, MD; G. Patrick Clagett, MD; Graham F. Pineo, MD, FCCP; Clifford W. Colwell, MD^§; Frederick A. Anderson, Jr., PhD^|| and H. Brownell Wheeler, MD

^* Dr. Geerts receives research funding from Aventis Pharma, Pharmacia & Upjohn. Dr. Heit receives research funding from AstraZeneca, Aventis Pharma, Corvas, DuPont Pharma, and Wyeth-Ayerst. Dr. Pineo receives research funding from DuPont Pharma, Emesphere Technologies, Leo Pharma, and Pharmacia & Upjohn. Dr. Pineo also serves on the advisory boards for Pharmacia & Upjohn and DuPont Pharma. ^§ Dr. Colwell receives research funding from and serves as consultant for Rhone-Poulenc Rorer, Pharmacia & Upjohn, and AstraZeneca. ^|| Dr. Anderson receives research funding from and serves as consultant for Aventis Pharma.

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.

	Introduction

TOP Introduction Why isn’t... Risk Factor Stratification Important Issues Related to... General, Gynecologic, and... Orthopedic Surgery Neurosurgery, Trauma, Acute... Burns Medical Conditions Prophylaxis Implementation... Recommendations General, Gynecologic, and... Major Orthopedic Surgery Neurosurgery, Trauma, and Acute... Medical Conditions References

 
This chapter reviews the literature related to the risks of VTE and its prevention. For each patient group, literature searches have been conducted and a priori criteria for inclusion of studies have been applied to derive quantitative estimates of the baseline risks of thromboembolism and the efficacy of each of the prophylaxis interventions (Table 1) . In the summary tables, the rates of deep vein thrombosis have been pooled from the eligible trials for each intervention and then compared with the rate among pooled, untreated, or placebo-treated control patients to determine the reduction in relative risk. Because comparisons among the interventions are indirect, the results of this pooling analysis provide an approximate guide to the relative efficacy of various prophylactic strategies. The final recommendations are based on the results of our pooled data as well as major randomized trials and/or formal, published meta-analyses. Although the recommendations are evidence-based, where possible, practical suggestions for prophylaxis are provided, particularly in situations where the evidence is inadequate.

An alternative to prophylaxis would be the use of serial surveillance tests such as duplex ultrasonography in high-risk patients.^{5 6} This approach is expensive and can be applied only to limited numbers of patients at risk. In addition, noninvasive screening tests, such as impedance plethysmography or duplex ultrasonography, have only moderate sensitivity and positive predictive value when used in asymptomatic, high-risk patients such as those undergoing major orthopedic surgery.^{7 8 9 10 11} Routine screening has also not been demonstrated to reduce the frequency of clinically important outcomes, such as symptomatic VTE or fatal PE. Broad application of effective methods of prophylaxis has been more cost-effective and is probably safer than selective, intensive surveillance.^{12 13 14 15 16 17 18 19 20 21}

Despite overwhelming evidence of the efficacy of an assortment of prophylactic modalities, surveys conducted in the United States,^{22 23 24} Canada,²⁵ the United Kingdom,^{26 27 28 29 30} Sweden,³¹ Switzerland,³² Spain,³³ and Australia/New Zealand^{34 35} document wide practice variations among physicians, with 28 to 100% of respondents indicating that they routinely used prophylaxis. In a random survey of fellows of the American College of Surgeons, 86% claimed they used prophylaxis in 1993,²³ this proportion rising to 96% by 1997.³⁶ However, a US study of 2,000 patients, hospitalized at 16 acute-care hospitals, showed that only one third of these patients actually received prophylaxis despite the presence of multiple risk factors for VTE.³⁷ Use of prophylaxis was higher in teaching than in nonteaching hospitals. A records review of patients aged 65 years or older and undergoing abdominal or thoracic surgery at 20 Oklahoma hospitals showed that prophylaxis was used in only 38%.³⁸ Of patients considered to be at very high risk, with multiple risk factors for VTE, only 39% received prophylaxis, and one third of these received inappropriate prophylaxis according to published guidelines. In 1996, a Scottish study, entitled "Still Missing the Boat With Fatal Pulmonary Embolism," documented fatal PE in surgical patients during a 1-year period.³⁹ Fifty-six percent of the patients who died of PE did not receive prophylaxis despite having major risk factors and no contraindications to standard antithrombotic regimens.

	Why isn’t thromboprophylaxis used more widely?

TOP Introduction Why isn’t... Risk Factor Stratification Important Issues Related to... General, Gynecologic, and... Orthopedic Surgery Neurosurgery, Trauma, Acute... Burns Medical Conditions Prophylaxis Implementation... Recommendations General, Gynecologic, and... Major Orthopedic Surgery Neurosurgery, Trauma, and Acute... Medical Conditions References

 
Many physicians believe that the overall incidence of VTE among hospitalized and postoperative patients has decreased over the past decades, to a point where the incidence is too low to consider prophylaxis. These physicians frequently cite informal, retrospective surveys of their own clinical services (or their personal experience) and the rare occurrence of fatal PE diagnosed by autopsy at their hospital to bolster this argument. In fact, the incidence of VTE may have declined in recent years,⁴⁰ and this probably reflects the success of prophylaxis strategies as well as other aspects of surgical and postoperative care.^{41 42 43} Even so, the incidence of this preventable condition remains too high; current estimates of the incidence of fatal PE, based on hospital discharge data, suggest the need for even wider application of prophylaxis.⁴⁴ Furthermore, the difficulty in establishing the antemortem diagnosis of PE is very common as is the low rate of autopsy in the United States. Data from countries where autopsies are carried out more commonly indicate that PE remains a significant problem.^{42 44 45 46} A 25-year population-based study from the Rochester Epidemiology Project documents that, while the incidence of PE has decreased during this period, the incidence of DVT has remained unchanged for men and is increasing for older women.⁴⁷ Most epidemiologic studies document that the elderly are particularly vulnerable to PE.^{46 47} With the increasing age of the population, VTE will become an escalating public health problem.

Another reason for failure to use prophylaxis, especially in surgical patients, is the concern about bleeding complications from anticoagulants. Countering this argument are abundant data from meta-analyses and placebo-controlled, double-blind, randomized trials that demonstrate either no increase or small increases in the absolute rates of major bleeding with the use of low-dose unfractionated heparin (LDUH) or low-molecular-weight heparin (LMWH).^{48 49 50 51 52 53 54} Although wound hematomas are seen more frequently with these agents,^{48 51} (and this may potentially increase the risk of wound infection, dehiscence, and infection of a prosthetic device placed at the time of operation), avoidance of LDUH or LMWH cannot generally be justified on these grounds alone. Alternatively, mechanical methods of prophylaxis carry no bleeding risk and have been efficacious in some patient groups.⁴⁸ Heparin-induced thrombocytopenia is also a potential concern with widespread use of heparin preparations.⁵⁵ The rate of thrombocytopenia with prophylactic use of heparin is 1 to 5%, and the incidence of clinically overt vascular thrombosis in postoperative patients with heparin-induced thrombocytopenia is approximately 50%.^{55 56} LMWHs are much less likely to produce heparin-induced thrombocytopenia than unfractionated heparin.⁵⁶ The costs of thromboprophylaxis have also been used as an argument against its wider use; however, the studies addressing this issue have uniformly concluded that broad application of prophylaxis is highly cost-effective.^{12 13 14 15 16 17 18 19}

The final major reason for not using prophylaxis has to do with subjective perceptions of the magnitude of the problem and the effects of prophylaxis in individual practices. Because VTE is most often clinically silent, the occurrence of overt VTE among an individual physician’s patients is perceived as rare.⁵⁷ For example, extrapolation of data from meta-analyses suggests that fatal PE occurs in 0.5 to 0.8% of unprotected patients over the age of 40 years undergoing major abdominal surgery and, in many of these, the diagnosis and cause of death would not be known.^{14 48 49} Similarly, although postoperative proximal DVT is present in 6 to 7% of general surgery patients, the majority do not have clinical manifestations and therefore would not be detected. As a consequence, a busy surgeon whose practice consists of a high volume of major abdominal surgery may not perceive VTE as a significant problem. More importantly, this physician would not be aware of a reduction in the incidence of fatal PE from 0.7 to 0.2% in his or her own practice that has been found in meta-analyses with the use of LDUH, for example.^{48 49} Thus, from an individual practice perspective, it is difficult to appreciate the effectiveness of prophylaxis, whereas failures (patients developing clinically overt VTE despite prophylaxis) are readily apparent. In addition, bleeding complications are highly visible, not easily forgotten, and frequently attributed, inappropriately, to the use of prophylaxis.

	Risk Factor Stratification

TOP Introduction Why isn’t... Risk Factor Stratification Important Issues Related to... General, Gynecologic, and... Orthopedic Surgery Neurosurgery, Trauma, Acute... Burns Medical Conditions Prophylaxis Implementation... Recommendations General, Gynecologic, and... Major Orthopedic Surgery Neurosurgery, Trauma, and Acute... Medical Conditions References

 
Knowledge of specific risk factors in patient groups or in individual patients forms the basis for the appropriate use of prophylaxis. Clinical risk factors include the following: increasing age; prolonged immobility, stroke, or paralysis; previous VTE; cancer and its treatment; major surgery (particularly operations involving the abdomen, pelvis, and lower extremities); trauma (especially fractures of the pelvis, hip, or leg); obesity; varicose veins; cardiac dysfunction; indwelling central venous catheters; inflammatory bowel disease; nephrotic syndrome; and pregnancy or estrogen use.^{58 59 60 61 62 63} These risk factors are present, often in combination, in a high proportion of hospitalized patients.⁶⁴ For surgical patients, the incidence of DVT is affected by the preexisting factors just listed and by factors related to the procedure itself, including the site, technique, and duration of the procedure, the type of anesthetic, the presence of infection, and the degree of postoperative immobilization. The role of congenital and acquired thrombophilic disorders (hypercoagulable states) in potentiating the risk of VTE associated with clinical risk factors (especially hospitalization or surgery) remains to be clarified. The thrombophilic abnormalities include the following: activated protein C resistance (factor V Leiden); prothrombin variant 20210A; antiphospholipid antibodies (lupus anticoagulant and anticardiolipin antibody); deficiency or dysfunction of antithrombin, protein C, protein S, or heparin cofactor II; dysfibrinogenemia; decreased levels of plasminogen and plasminogen activators; heparin-induced thrombocytopenia; hyperhomocystinemia; and myeloproliferative disorders such as polycythemia vera and primary thrombocytosis.^{65 66 67 68}

In many patients, multiple risk factors may be present, and the risks are cumulative.^{69 70} For example, elderly patients with hip fractures undergoing major orthopedic operations and who remain immobile in bed after operation are among the most susceptible to fatal PE. Formal risk assessment models for DVT have been proposed for surgical patients.^{71 72 73 74 75 76 77} Awareness of the clinical settings in which the risk has been defined by epidemiologic studies is also important in the successful application of prophylaxis recommendations (Table 2 ). For example, the patients at greatest risk for VTE are those undergoing major lower extremity orthopedic surgery and those who experience major trauma or spinal cord injury.

	Important Issues Related to Thromboprophylaxis Data

TOP Introduction Why isn’t... Risk Factor Stratification Important Issues Related to... General, Gynecologic, and... Orthopedic Surgery Neurosurgery, Trauma, Acute... Burns Medical Conditions Prophylaxis Implementation... Recommendations General, Gynecologic, and... Major Orthopedic Surgery Neurosurgery, Trauma, and Acute... Medical Conditions References

Appropriate End Points for Studies of DVT Prophylaxis:
Physicians differ in their views on the appropriate end points for studies of DVT prophylaxis. Some believe that very sensitive and specific diagnostic tests for all thromboembolic activity are essential. These outcomes are contrast venography for high-risk patients and fibrinogen leg scanning for moderate-risk patients. Others consider that evidence of reduction in deaths from all causes is required to convince them that an intervention is of benefit. Both of these approaches have limitations. The majority of the thrombi detected by sensitive screening methods for DVT are not clinically relevant (although only a small amount of data allows us to predict which thrombi will resolve and which will produce important adverse effects). However, insistence on death as the exclusive outcome dismisses the significant burden of disease due to symptomatic thromboembolic events as well as the cost inefficiency associated with the investigation and treatment of these complications. We suggest a middle ground based on large trials that use a clinically important VTE outcome, consisting of a composite of fatal PE, symptomatic, proven DVT or PE, and asymptomatic proximal DVT. These larger trials should be performed once smaller studies using an accurate test for all DVT have demonstrated the biological efficacy of the intervention.

Limitations of DVT Screening Methods
Each of the DVT screening methods has limitations. Fibrinogen leg scanning, also called the fibrinogen uptake test (FUT), lacks specificity and sensitivity^{78 79 80} ; duplex ultrasonography has poor sensitivity as a screening test in asymptomatic patients^{8 9 10 11} ; and venography is associated with a significant rate of nondiagnostic studies, is no longer widely available, and the clinical relevance of many of the thrombi detected is questioned. Despite these limitations, the relative risk reductions when two prophylaxis choices are compared using these outcome measures are likely to be valid as long as systematic bias has been eliminated.⁸¹

Mechanical Methods of Prophylaxis:
Special caution should specifically be exercised when interpreting the risk reductions ascribed to mechanical methods of prophylaxis for three reasons. Most trials have not been able to blind the mechanical devices, leading to the potential for diagnostic suspicion bias. If fibrinogen leg scanning was the DVT screening method, the known 10 to 30% false positive rate of the FUT might have been reduced by the mechanical prophylaxis but not by the alternative option.⁸² Finally, because of relatively poor compliance with all mechanical options, they may well not perform as well in routine clinical practice as in research studies where major efforts are made to optimize proper use.

Results May Not Apply to All Patients:
Because most studies have excluded the patients at highest risk for both thromboembolic and adverse outcomes, the results may not apply to all patients, especially those with previous history of VTE, or to patients with a greater-than-average bleeding potential.

Prophylaxis Decisions for an Individual Patient:
The prophylaxis recommendations contained herein are made for groups of patients, for whom the benefits appear to outweigh the risks. However, prophylaxis decisions for an individual patient are best made by combining knowledge of the literature (including the group recommendations provided herein and elsewhere) with clinical judgment (including detailed knowledge of that particular patient’s unique risks for thrombosis, the potential for adverse consequences due to the prophylaxis, and the availability of various prophylaxis options locally). The recommendations that are best for the group may not be best for the individual.⁸³

Antithrombotic Drugs and Regional Anesthesia:
Perispinal hematoma after neuraxial blockade (spinal or epidural anesthesia or epidural analgesia) is a rare complication of anticoagulant therapy or prophylaxis.^{84 85} Although rare, the seriousness of the complication mandates cautious use of antithrombotic medication in patients having neuraxial blockade. A 1997 US Food and Drug Administration Public Health Advisory called attention to safety reports describing 43 US patients who had developed perispinal hematoma after receiving the LMWH enoxaparin concurrently with spinal/epidural anesthesia.^{86 87} Many of these patients suffered neurologic impairment, including permanent paralysis, despite decompressive laminectomy in 65%. The median age was 78 years (range, 28 to 90), and 78% of the patients were women. Some patients had preexisting spinal abnormalities, and a third received additional hemostasis-inhibiting medications. Nearly 90% of these complications occurred in patients receiving enoxaparin as prophylaxis after surgery, primarily total knee or hip replacement or spinal surgery. Factors suspected of predisposing patients to perispinal hematoma include the presence of an underlying hemostatic disorder, traumatic needle or catheter insertion, repeated insertion attempts or blood return, catheter insertion or removal in the presence of significant levels of anticoagulant, use of continuous epidural catheters, anticoagulant dosage, concurrent administration of medications known to increase bleeding, vertebral column abnormalities, older age, and female gender.^{84 85 87} Unfortunately, the prevalence of this problem and the predictive value of potential risk factors are, at present, unknown. The problem has also been reported with LDUH, although with apparent lower frequency. Therefore, the benefit vs risk of any anticoagulant prophylaxis or therapy for patients with spinal/epidural anesthesia or analgesia is difficult to assess.

Critical reviews of this problem provide guidelines for LMWH use in patients with spinal/epidural anesthetic interventions.^{85 88 89} We believe that neuraxial blockade and anticoagulant thromboprophylaxis, including LMWHs, can generally be used concomitantly. The following recommendations may improve the safety of neuraxial blockade in patients who have received or will receive anticoagulant prophylaxis: (1) neuraxial blockade should generally be avoided in patients with a clinical bleeding disorder; (2) in patients receiving drugs that may impair hemostasis (eg, aspirin, other platelet inhibitors, or anticoagulants), insertion of the spinal needle should be delayed until the anticoagulant effect of the medication is minimal (usually at least 8 to 12 h after a prophylactic LMWH or heparin injection); (3) anticoagulant prophylaxis should be avoided or delayed if there is a hemorrhagic aspirate ("bloody tap") during the initial spinal needle placement; (4) removal of epidural catheters should be done when the anticoagulant effect is at a minimum (usually just before the next scheduled subcutaneous [SC] injection); and (5) anticoagulant prophylaxis should be delayed for at least 2 h after spinal needle placement or catheter removal. All patients should be monitored carefully and frequently for the new onset of back pain and for symptoms or signs of cord compression (eg, progression of lower extremity numbness or weakness, bowel or bladder dysfunction). For patients in whom spinal hematoma is suspected, diagnostic imaging and definitive surgical therapy must be performed as rapidly as possible to reduce the risk of permanent paresis.

The sections that follow are based primarily on the hospital services to which patients are admitted. In each patient category, the risks of VTE and the effective methods of prophylaxis are detailed, if known. For most patient groups, sufficient numbers of randomized clinical trials are available to allow strong recommendations (grade 1A or 1B) to be made with regard to the benefits and risks of methods to prevent VTE. Standard antithrombotic regimens shown to be effective are summarized in Table 3 .

	General, Gynecologic, and Urologic Surgery

TOP Introduction Why isn’t... Risk Factor Stratification Important Issues Related to... General, Gynecologic, and... Orthopedic Surgery Neurosurgery, Trauma, Acute... Burns Medical Conditions Prophylaxis Implementation... Recommendations General, Gynecologic, and... Major Orthopedic Surgery Neurosurgery, Trauma, and Acute... Medical Conditions References

The advantages and disadvantages of LMWH in general surgery have been clarified by a number of large trials, as well as by meta-analyses in which LMWH and LDUH were compared.^{50 51 147 148 149 150 151 152 153 154 155 156 157 158 159 160} On balance, LMWH and LDUH appear to be equally efficacious in preventing DVT in general surgery patients. Some studies have reported significantly fewer wound hematomas and bleeding complications with LMWH,^{51 157 159} while other well-designed trials have shown that LMWH causes more bleeding than LDUH.^{149 153 160} The discrepant findings appear to be related to dosage; there is a clear dose-response effect of LMWH on bleeding complications (and probably also on the efficacy of prophylaxis). Higher doses of LMWH ( > 3,400 anti-Xa units daily) in comparison to LDUH (5,000 U bid or tid) are associated with more bleeding.⁵³ In contrast, lower doses of LMWH (< 3,400 anti-Xa units daily) are equivalent to LDUH in preventing VTE in moderate-risk patients and have a lower rate of bleeding complications.⁵³ While one meta-analysis could not discern superior efficacy of higher doses of LMWH,⁵³ individual studies in high-risk general surgery patients suggest that this may be the case.^{158 172 173} One distinct advantage of LMWH is that it can be administered once daily. LMWH is also less likely to cause heparin-induced thrombocytopenia and thrombosis than standard heparin preparations.⁵⁶ Optimal timing for the commencement of LMWH therapy (preoperatively or postoperatively) has been the subject of considerable interest. In orthopedic patients, anticoagulant treatment is often started 12 to 24 h after operation because of fear of bleeding and for convenience. In general surgery patients, there appear to be no adverse consequences of giving the first dose of LMWH ( < 3,400 U) 2 h before operation,¹⁷⁷ and there may be an additional benefit in preventing DVT from developing during surgery or in the immediate postoperative period. When higher doses of LMWH are used in high-risk general surgery patients, treatment with the drug should generally be commenced 10 to 12 h before operation to avoid excessive intraoperative bleeding.

Given the approximate equivalence in efficacy and safety of LDUH and LMWH in general surgery patients, cost becomes an important determinant in the choice between these drugs. In North America, LMWHs cost 2 to 10 times more than LDUH, and the cost-effectiveness analyses performed in abdominal and colorectal surgery patients concluded that prophylaxis with LDUH was more economical.^{160 179} In countries where LMWHs are less expensive, these agents may be equivalent in overall costs and more appealing because of once daily administration.^{19 180}

Intermittent pneumatic compression (IPC) is an attractive method of prophylaxis because there is no risk of hemorrhagic complications. However, IPC has not been studied as thoroughly as other methods in general surgery. Several small studies have demonstrated that IPC is effective in reducing DVT in general surgery patients and in surgical patients with malignant disease.^{48 129 140} In trials comparing IPC with LDUH, both agents produced similar reductions in DVT.^{161 162 171} It is not proven that IPC prevents PE (or even proximal DVT) in general surgery patients. Intermittent plantar compression, using the venous foot pump, produces hemodynamic effects on lower extremity emptying similar to that of IPC and, like IPC, it also stimulates fibrinolytic activity.¹⁸¹ To our knowledge, there are no trials of these devices in general surgery patients.

Graded compression elastic stockings (ES) reduce the incidence of leg DVT¹⁸² and enhance the protection provided by LDUH, but too few data are available to assess their effect on proximal DVT and PE. Patients with malignant disease and other high-risk general surgical conditions have not been evaluated in sufficient numbers to allow firm conclusions with regard to the efficacy of ES in these clinical settings. In some of the randomized trials, high-risk patients were specifically excluded.^{119 120} Further clinical trials are needed to assess the effectiveness of ES in such patients. Another limitation is that some patients cannot effectively wear ES because of unusual limb size or shape.

Combining ES with other prophylactic agents, such as LDUH, appears to give better protection against VTE than either approach alone.^{163 183} ES counteract venous stasis and augment venous return during abdominal insufflation for laparoscopic procedures.^{184 185} A recent uncontrolled study demonstrated a 2% risk of DVT as detected by duplex ultrasonography in patients undergoing laparoscopic or minilaparotomy cholecystectomy when LMWH, intraoperative IPC, and ES were combined.¹⁸⁶

Because of its low expense, ease of administration, and few side effects, aspirin would appear to be an ideal antithrombotic agent to prevent VTE. However, aspirin has generally been found to be ineffective in preventing VTE in general surgery patients, and we do not recommend it as an appropriate strategy.⁷⁵ This view has been challenged by the Antiplatelet Trialists’ Collaboration meta-analysis, which concluded that perioperative antiplatelet treatment reduced the incidence of DVT in general surgery patients by 37% and PE by 71% in comparison to untreated control subjects.¹⁸⁷ These reductions were highly significant, and similar effects were also reported in patients undergoing orthopedic and other operations. However, the Antiplatelet Trialists’ Collaboration group pooled > 30 antiplatelet trials of variable scientific design.^{188 189} Most individual trials demonstrate no significant benefit of aspirin or they show that aspirin is less effective than other agents.^{110 122 123 124}

Despite the paucity of evidence, warfarin, given in full therapeutic doses, may be effective in preventing extensive DVT in general surgery patients.¹¹⁵ However, the onset of action of warfarin is delayed, the treatment is cumbersome because it requires frequent laboratory monitoring, and it is subject to bleeding complications if not closely monitored. Because of these shortcomings and the availability of other effective options, there is little rationale for using warfarin in general surgery patients.

An appropriate preventive strategy in general surgery takes into account the risk of VTE, the effectiveness of the various agents, and the expense and possible complications incurred by their use (Table 2) .⁷⁵ In low-risk patients undergoing minor or relatively short operations, who are < 40 years of age and have no additional risk factors, no specific prophylaxis other than early ambulation is necessary. Two large-scale studies document a near zero risk for the development of clinical VTE after minor procedures in low-risk patients.^{190 191} In moderate-risk patients who are > 40 years of age or who are undergoing major operations, but who have no additional clinical risk factors, LDUH given every 12 h, LMWH once daily (< 3,400 anti-Xa U), or properly used ES should be sufficient. IPC would be a reasonable alternative to these agents. In patients > 40 years undergoing major surgery with additional risk factors, several effective prophylactic methods are available. LDUH given every 8 or 12 h and once-daily LMWH are effective. IPC would also be a consideration, especially if the patient is particularly prone to bleeding. Adding ES to any of these methods may give additional protection. In general surgery patients with multiple risk factors, combining the most effective pharmacologic methods with IPC or ES should offer excellent protection. Higher daily doses of LMWH (> 3,400 U), as is often used in orthopedic surgery, would also be appropriate.

The issue of prophylaxis beyond the period of hospitalization was addressed in a single small, randomized study of high-risk patients undergoing major abdominal or thoracic surgery.¹⁹² Prolonged prophylaxis with LMWH for 3 weeks after hospital discharge did not significantly reduce the incidence of DVT as assessed by bilateral venography performed 4 weeks after surgery, compared with 1 week of in-hospital LMWH (5% vs 10%). However, a total of only 118 patients had adequate venography. A cost-effectiveness analysis, based on event rates from the literature, concluded that postdischarge prophylaxis of general surgery patients was effective, but the marginal costs were too high to warrant its routine use.¹⁹³ The issue of duration of thromboprophylaxis in general surgery must now be reevaluated in the context of current short lengths of hospital stay.

Gynecologic Surgery
VTE is also an important and potentially preventable complication following gynecologic surgery.^{194 195 196 197 198} The overall incidence of DVT is comparable to or slightly lower than that associated with general surgery.¹⁹⁹ Using the FUT as the primary outcome measure, the reported frequency of postoperative DVT in 19 studies that included 2,268 patients who underwent gynecologic surgery without prophylaxis varied between 4% and 38%, with an average of 16%.^{93 96 107 115 119 127 134 200 201 202 203 204 205 206 207 208 209 210 211} Fatal PE has been reported in 0.4% of a pooled sample, including > 1,000 unprotected patients.^{96 107 119 133 202 208 209} The factors that appear to increase the thromboembolic risk following gynecologic surgery include older age, previous VTE, surgery for cancer, and abdominal (vs vaginal) procedure. Gynecologic oncology patients, in particular, have a substantially increased DVT risk because many of these patients are elderly; they all have cancer; in some there may be compression of major veins by a pelvic mass; they are prone to venous intimal injury during the procedure, especially when pelvic lymph node dissection is performed; the procedures are frequently lengthy; residual tumor may be left behind; postoperative mobility is often impaired; and chemotherapy itself is thrombogenic. As in other surgical patients, although thrombi generally begin to form at or shortly after surgery,²⁰⁸ the majority of symptomatic events occur after hospital discharge.²¹² Despite changes in surgical and postoperative care and the use of prophylaxis, few prospective studies have been carried out over the past 15 years. Therefore, contemporary data related to the risks and prevention of VTE in this patient group are lacking.²¹²

Pooling of the rates of fatal PE in prospective studies of 7,000 gynecologic surgery patients demonstrates a 75% risk reduction with the use of thromboprophylaxis (from 0.4 to 0.1%). The results of randomized trials of prophylaxis on DVT rates in gynecologic surgery patients are displayed in Table 6 .

The strongest evidence that thromboprophylaxis is of benefit in gynecologic surgery has been provided for the use of LDUH. In six randomized trials with untreated control groups, the relative risk reduction in DVT with LDUH treatment was 64% (20% vs 7%).^{93 96 107 115 203 209} Patients having surgery for gynecologic cancers derive less protection from twice daily administration of LDUH than those with benign disease,^{96 217} while a regimen of LDUH given three times daily appears to be more effective in these patients.^{96 209 215} Increased bleeding complications have been described in some studies using LDUH,^{215 220} but not in others.²⁰⁹

When compared with LDUH, aspirin and dextran have an efficacy rate 2 to 4 times lower in gynecologic surgery patients and are not recommended.^{107 203 216 217} Treatment with oral anticoagulants in full doses or in mini doses, started at least a week before surgery, has been more efficacious than no prophylaxis in three small studies,^{115 210 211} but LDUH is at least as effective and considerably easier to use.¹¹⁵ To the best of our knowledge, there are no trials using LMWH that meet the inclusion criteria in Table 1 , although LMWH appears to provide protection comparable to LDUH when either symptomatic VTE or screening with impedance plethysmography is employed.^{221 222 223 224 225 226} In an uncontrolled case series of 2,030 patients who had major gynecologic surgery and who were given enoxaparin 20 mg once daily, there was one fatal PE, and only 7 patients (0.3%) had symptomatic VTE.²²⁷

The risk classification and prophylaxis recommendations in Table 2 are applicable to gynecologic surgery.^{196 197 199} Patients who are otherwise well and who undergo brief procedures probably do not require any specific prophylaxis, but they should be encouraged to mobilize early after surgery. For patients having major gynecologic procedures for benign disease without additional risk factors, administration of LDUH twice daily is recommended. Alternatives include treatment once daily with LMWH or intraoperative IPC continued for at least several days after surgery. For higher-risk patients, one of the following options is recommended: LDUH + ES or IPC, LDUH three times daily, or LMWH given in daily doses of at least 3,400 anti-Xa U. An unresolved issue is the duration of antithrombotic therapy in gynecologic oncology patients. A recent study followed a large cohort of gynecologic cancer patients with serial IPGs postoperatively and during subsequent courses of chemotherapy.²²⁸ The postoperative proximal DVT rate was 15%, but this increased to 20 to 30% when the events during follow-up were also counted. The occurrence of these thrombi predicted a sixfold increased risk of death during follow-up.

Urologic Surgery
Thromboembolic events are considered the most important nonsurgical complication of major urologic procedures.^{229 230 231} Because most of the epidemiologic data were derived 10 to 30 years ago, changes in surgical care, more aggressive mobilization, and possibly greater use of prophylaxis have apparently resulted in decreased rates of VTE over time.^{232 233} However, 1 to 5% of contemporary patients undergoing major urosurgery experience symptomatic VTE, and fatal PE is seen occasionally (risk 1/500).^{232 233 234 235 236 237 238 239 240} Factors that have been demonstrated to increase the risk of DVT in these patients include open (vs transurethral) procedures, malignancy, increased age, general (vs regional) anesthetic, and duration of the procedure.

Over the past decade, to our knowledge, there have been no published studies in urology that meet the methodologic criteria in Table 1 , and the optimal approach to thromboprophylaxis is not known for these patients.²⁴¹ LDUH and LMWH probably have similar efficacy in urology as in general surgery.^{49 103 118 227 231 240} However, concerns have been raised about the potential for increased rates of pelvic hematoma and lymphocele in patients receiving anticoagulant prophylaxis for open urologic procedures.^{231 233 240} Use of ES or IPC is likely to be efficacious in urosurgery,^{97 236 242 243} but the high costs of IPC have been raised as a problem with this method.²⁴⁴ It is also possible that the addition of IPC to inexpensive ES may not provide additional protection in these patients.^{236 242} However, combining mechanical and pharmacologic prophylaxis may be more effective than either alone but will substantially increase the costs.^{112 231 240}

For patients undergoing transurethral prostatectomy, the risks of VTE are low,^{49 103 227} and there may be increased risk of bleeding with use of perioperative LDUH or LMWH.^{245 246 247} Early postoperative mobilization is probably the only intervention warranted in these and other low-risk urosurgery patients. Routine prophylaxis is recommended for more extensive, open procedures, including radical prostatectomy, cystectomy, or nephrectomy. Until further data become available, the options to consider for these patients include LDUH, ES, IPC, LMWH, and combinations of mechanical and pharmacologic methods. For patients at particularly high risk, commencing treatment with ES plus or minus IPC just prior to surgery and then adding LDUH (or LMWH) postoperatively should be considered, although this approach has not been formally evaluated in urology patients (to our knowledge).

	Orthopedic Surgery

TOP Introduction Why isn’t... Risk Factor Stratification Important Issues Related to... General, Gynecologic, and... Orthopedic Surgery Neurosurgery, Trauma, Acute... Burns Medical Conditions Prophylaxis Implementation... Recommendations General, Gynecologic, and... Major Orthopedic Surgery Neurosurgery, Trauma, and Acute... Medical Conditions References

 
Clinical trials and cohort studies have provided a clearer picture of the natural history of acute VTE associated with major orthopedic surgery of the lower extremity and have also provided considerable information to guide decisions about prophylaxis. Based on the results of contrast venography, performed on either control patients or patients randomized to receive placebo, the prevalence of total DVT at 7 to 14 days after total hip replacement (THR), total knee replacement (TKR), and hip fracture surgery is about 50 to 60% (Table 7 ),^{248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272} with proximal DVT rates of about 25%, 15 to 20%, and 30%, respectively. While the operated-on leg is most commonly affected, the nonoperated-on leg is also affected in about 20% of THR patients^{250 274 275 276 277 278 279} and in about 14% of TKR patients.²⁸⁰ The incidence of asymptomatic PE is less certain. Intraoperative transesophageal echocardiography shows frequent "debris" transiting the right side of the heart, particularly during reaming of the bone.^{281 282} This debris, which includes both fat and thromboemboli, often causes transient hypoxemia and pulmonary hypertension; however, serious clinical sequelae are uncommon. In studies in which a ventilation-perfusion lung scan was routinely performed, about 7 to 11% of THR and TKR patients had a high-probability scan within 7 to 14 days after surgery.^{252 259 261 283 284} New DVT and PE after hospital discharge are also common. Venography studies show that, without postdischarge prophylaxis, 10 to 20% of patients develop new evidence of DVT within 4 to 5 weeks after hospital discharge,^{284 285 286} and about 6% develop an intermediate- or high-probability lung scan.²⁸⁴

Together, these data suggest the following hypothesis regarding the natural history of VTE after major orthopedic surgery. Asymptomatic VTE (including proximal DVT and even PE) is common and, in the absence of prophylaxis, affects at least half of these patients. The majority of these thrombi resolve spontaneously. For certain patients, however, the persistence of venous injury, stasis due to prolonged immobility,²⁹⁴ an impaired natural anticoagulant²⁹⁵ or fibrinolytic system, or some as yet unidentified factor, allows a thrombus to propagate and to become symptomatic due to either venous occlusion or embolization. At present, our ability to identify these high-risk patients is limited, and future research should be directed to determining the genetic, clinical, and biochemical characteristics that predispose to the development of clinically important postoperative VTE. Until we are able to stratify patients according to their individual risk and then target prophylaxis to those at highest risk, primary prophylaxis should be provided to all patients undergoing major orthopedic surgery of the lower extremity. While most DVT detected by venography will remain asymptomatic and will resolve without treatment, thrombosis detected by venography remains a credible outcome measure for comparing the efficacy of different prophylaxis regimens. Consequently, we have confined our review to English-language clinical trials that required either mandatory postoperative venography of the operated-on leg (or both legs) or objectively confirmed symptomatic VTE for determination of efficacy. Since we cannot predict which asymptomatic DVT will eventually become symptomatic,^{296 297 298 299 300} we have analyzed the total DVT rates (proximal plus distal DVT). We report the pooled venography results (including 95% confidence intervals [CIs] and relative risk reductions) by type of surgery (THR, TKR, or hip fracture surgery) to allow cross-trial comparisons of different prophylaxis agents and regimens. Only results from single-modality prophylaxis regimens (excluding graded elastic compression stockings) are included. Finally, the benefits of any prophylaxis regimen should be weighed against the costs, including those resulting from bleeding complications, as well as the costs associated with failed prophylaxis (eg, VTE and death). This comparison is best performed using a formal cost-effectiveness analysis.³⁰¹ Although we report cost-effectiveness studies where available, they should be interpreted with caution, as most used risk reduction in asymptomatic DVT by venography to determine the potential benefit derived from each prophylaxis regimen.

Elective THR Surgery
Withholding primary prophylaxis in favor of case-finding by serial noninvasive screening for asymptomatic DVT is problematic in this patient population because the commonly available noninvasive tests (impedance plethysmography or compression or color duplex ultrasonography) are insensitive for asymptomatic calf and proximal DVT.^{7 8 9 302 303 304 305 306} Moreover, clinical trials and cohort studies have found that a strategy of screening for proximal DVT with predischarge color duplex ultrasonography was ineffective.^{287 307} While a similar strategy using predischarge venography appeared to be cost-effective in a single study,³⁰⁰ routine venography is not widely available or generally acceptable. Radioisotope-based imaging of asymptomatic thrombus has not been shown to be beneficial in large studies.³⁰⁸ Consequently, primary prophylaxis is recommended for all THR patients (Table 8 ).

Inferior vena cava (IVC) filter placement has been suggested as a prophylaxis option for patients at extremely high risk for both postoperative VTE and bleeding.^{361 362 363} However, we are not aware of any randomized trials of prophylactic IVC filter insertion or of any studies that address the value of filters when added to recommended prophylaxis options. In a treatment study, patients with acute DVT who were judged to be at high risk for subsequent PE were randomly assigned to IVC filter placement or no filter placement in addition to concurrent anticoagulation.³⁶⁴ The incidence of subsequent PE (symptomatic plus asymptomatic) was significantly reduced in the short-term among patients receiving an IVC filter. However, mortality was not reduced in the filter group, and filter patients had significantly more recurrent DVTs on follow-up. Extrapolating these data to high-risk orthopedic surgery patients, prophylactic IVC filter placement may reduce the immediate risk of post-operative PE at enormous cost, but it will increase the risk of future DVT.³⁶⁵ Based on these issues, we believe that placement of an IVC filter as prophylaxis should be discouraged.

A number of anticoagulant-based prophylaxis regimens for THR surgery have been studied (Table 8) . Although meta-analyses have shown LDUH⁴⁹ or aspirin¹⁸⁷ prophylaxis to be more effective than no prophylaxis, both are less effective than other prophylaxis regimens in high-risk patients.^{276 321 323 327 328} Among 4,088 hip and knee arthroplasty patients randomized to treatment with aspirin or placebo (plus or minus other thromboprophylactic measures), there was no benefit associated with aspirin use for either venous or arterial thromboembolic events.³⁶⁶ Preoperative LDUH followed by postoperative heparin, dose-adjusted to maintain the activated partial thromboplastin time at or just above the upper range of normal (adjusted-dose heparin), is safe and highly effective, and may be considered for patients at extremely high risk because of concomitant risk factors.^{275 323 343 346} However, most surgeons consider adjusted-dose heparin prophylaxis to be impractical for routine use.

Adjusted-dose oral anticoagulation (eg, warfarin sodium) is, generally, a safe and effective prophylaxis and has been adopted by many orthopedic surgeons in North America.^{25 253 329 333 334 335 367 368 369} Adjusted-dose warfarin has the potential advantage of allowing continued prophylaxis after hospital discharge (as long as the infrastructure is in place to do this effectively and safely). Oral anticoagulants should be administered at a dose sufficient to prolong the international normalized ratio (INR) to a target of 2.5 (range = 2.0 to 3.0). The initial oral anticoagulant dose should be administered either the evening prior to surgery or as soon after surgery as possible. However, even with early initiation of oral anticoagulant therapy, the INR usually does not reach the target range until at least the third postoperative day.^{334 337 370}

LMWH and heparinoids have been studied extensively and are highly effective and generally safe asVTE prophylaxis after THR (Table 8) . LMWH is more effective than LDUH,^{50 276 279 325} and is at least as effective as ,²⁷⁹ or superior²⁷⁵ to, adjusted-dose unfractionated heparin.

Two clinical trials that have compared LMWH to adjusted-dose warfarin prophylaxis found no difference in either total or proximal DVT prevalence.^{333 334} Among patients receiving LMWH prophylaxis, one trial showed a small increase in the number of bleeding complications,³³³ while the other study found greater blood loss.³³⁴ Another clinical trial found the total DVT prevalence to be significantly less with LMWH (started preoperatively) compared to adjusted-dose warfarin although, in this study, patients receiving LMWH prophylaxis had significantly greater bleeding at the operative site and greater transfusion requirements.³³⁷ Finally, a study comparing LMWH (started at half the daily dose, either within 2 h before surgery or at least 4 h after surgery) with warfarin therapy started postoperatively revealed a significant reduction in both total and proximal DVT rates with LMWH.³³⁹ The incidence of symptomatic, objectively documented DVT was also lower with preoperative LMWH, than with warfarin (1.5% vs 4.4%; p = 0.024).

Two meta-analyses of the various prophylaxis regimens concluded that LMWH was most effective, although the differences in efficacy between LMWH and either adjusted-dose warfarin or adjusted-dose heparin prophylaxis were small.^{371 372} When the results from the five large studies that directly compared adjusted oral anticoagulation with LMWH in THR were pooled, the DVT rates were 20.7% (256/1,238) in the oral anticoagulant groups and 13.7% (238/1,741) in the patients who received LMWH.^{333 334 335 337 339} The proximal DVT rates were 4.8% and 3.4%, respectively. The pooled rates for major bleeding (using somewhat different definitions in the five studies) were 3.3% in the oral anticoagulant patients and 5.3% in the LWMH groups. In a large, open-label clinical trial, THR patients were randomly assigned to in-hospital prophylaxis with either LMWH (enoxaparin 30 mg SC bid started postoperatively; N = 1,516) or adjusted-dose warfarin (INR = 2.0 to 3.0; N = 1,495).²⁸⁸ Symptomatic, objectively documented VTE was the primary efficacy end point. The mean duration of prophylaxis was 7.5 days for LMWH and 7.0 days for warfarin. The cumulative in-hospital incidence of symptomatic VTE was 0.3% among patients receiving LMWH, compared to 1.1% among patients receiving warfarin (p = 0.008). Major bleeding was seen in 1.2% of the LMWH patients and in 0.5% of the patients receiving warfarin (p = 0.055).

From these data, we conclude that LMWH is significantly more effective than warfarin in preventing asymptomatic and symptomatic in-hospital VTE. However, the risk of surgical site bleeding and wound hematoma is slightly greater with LMWH. These conclusions are consistent with the more rapid onset of anticoagulant activity with LMWH compared to warfarin. We suggest that the selection of LMWH or warfarin prophylaxis be made at the specific hospital level (and, on occasion, at the individual patient level) based on issues that include cost, convenience, availability of an infrastructure to provide safe oral anticoagulation, duration of planned prophylaxis, and potential bleeding and thrombosis risks. In a decision-analysis using Canadian health-care costs, LMWH was preferred over adjusted-dose warfarin anticoagulation.³⁷³ However, a recent analysis based on US health-care costs found adjusted-dose warfarin to be more cost-effective compared to LMWH.²¹

Three clinical trials have found treatment with SC recombinant hirudin (15 mg SC bid, initiated preoperatively) to be more effective than LDUH^{327 328} or LMWH,³⁴² with no difference in bleeding. While not approved for prophylaxis, recombinant hirudin (lepirudin, Refludan) is approved by the US Food and Drug Administration for therapy of heparin-induced thrombocytopenia.

Elective TKR Surgery
From the thromboembolism perspective, knee arthroplasty differs from THR in several important respects. Without prophylaxis, the total DVT rate is greater in TKR than in THR. The prophylaxis interventions, used successfully in THR, have significantly lower efficacy in TKR patients. Although major bleeding is not more common in TKR patients, awareness of and concerns about nonmajor bleeding and its potential consequences are greater after TKR. Finally, in TKR, LMWH clearly has greater efficacy than warfarin.

The results of four small studies suggest that IPC is effective prophylaxis in TKR patients^{331 374 375 376 377 378 379} (Table 9 ).These devices are most effective when they are applied either intraoperatively or immediately postoperatively and are worn continuously, at least until the patient is fully ambulatory. The utility of IPC devices is limited by poor compliance and patient intolerance, significant costs, and the inability to continue prophylaxis after hospital discharge. IPC may be useful as an in-hospital adjunct to anticoagulant-based prophylaxis regimens. The venous foot compression pump has been shown to be efficacious in two small studies in TKR patients.^{376 380} However, in two other trials, LMWH was considerably more efficacious than these devices.^{385 386} Continuous passive motion devices have not reduced the DVT incidence in TKR patients compared with routine physiotherapy alone.²⁶⁰

LMWH has been studied extensively and is safe and effective prophylaxis after TKR surgery.^{274 280 287 321 333 335 370 378 381}