Venous Thromboembolism (VTE)

Pregnancy. Warfarin can cause birth defects. Notify your doctor if you are pregnant.

Clinical Background
Clinical Problem and Current Dilemma
Deep venous thrombosis (DVT) and pulmonary embolism (PE) together comprise the spectrum of venous thromboembolic disease (VTE). VTE is one of the most frequent causes of hospitalization for adults and often complicates surgery and childbirth, carries significant risk of death and of long-term sequelae such as postphlebitic syndrome. Historically, prior to the widespread use of heparin, approximately 12% of all patients with clinically evident DVT died, most often from PE. Clinical findings are not adequate for diagnosis or exclusion. New imaging modalities are important, but their characteristics need to be understood and incorporated into cost-effective diagnostic strategies. Management of heparinization is variable. Over- and undershooting target levels is commonplace and extends hospital stays. Some patients are not able to receive warfarin, and some cannot receive any anticoagulation at all, complicating management of their VTE. The state of the art in managing VTE is changing rapidly with the introduction of low-molecular-weight heparin. LMWH may soon supplant UFH for most or all indications. Throughout this document DVT of the veins distal to the knee is not distinguished from proximal DVT. There has been an informal clinical tradition of regarding below-knee DVT as not requiring treatment, or being amenable to observation. However, studies of PE rates find that over
30% of distal DVTs embolize (compared to 50% of proximal ones), and symptomatic recurrence rates for untreated distal DVT exceed 30%. The risks posed by distal DVTs are lower than proximal DVT, but not greatly so, and not enough to merit less serious treatment.

Rationale for Recommendations
Diagnosis of Deep Venous Thrombosis
Clinical recognition of possible DVT.
The clinical diagnosis of DVT is challenging and characterized by uncertainty. DVT may be suspected in the settings listed under “Clinical situation” in Table 1, but is by no means limited to these settings. Typical symptoms and signs include swelling and tenderness of the calf, and Homan’s sign (slight pain at the back of the knee or calf when the ankle is slowly and gently dorsiflexed, with the knee bent). However, half of significant DVTs are without clinical symptoms or signs, so these may not be relied on for diagnosis. Superficial thrombophlebitis may closely resemble DVT, as may ruptured Baker’s cyst, gastrocnemius-soleus muscle injuries, and other conditions. The diagnosis cannot be made or excluded on clinical grounds, therefore threshold for testing should be low.

Testing for DVT.
The standard clinical practice for the diagnosis of deep venous thrombosis has become venous duplex imaging, with its most recent innovation being the use of color. The test characteristics of venous duplex imaging are presented in Table 4. The high negative predictive values (NPV) suggest that withholding anticoagulation on the basis of a negative study is appropriate. In the asymptomatic patient, there is wider range of positive predictive values (PPV). Below knee PPV for DVT diagnosis ranges as low as 75% to as high as 100%. NPVs remain good at the below-knee location, again suggesting that withholding anticoagulation on the basis of a negative study is appropriate. As noted in the table, color duplex is superior to grey scale and significant improvement in the sensitivity at the below
knee level has been found with the addition of color to duplex imaging. The use of venous duplex imaging for upper extremity deep venous thrombosis has also been documented. Venous duplex imaging is superior to indirect tests such as continuous wave doppler for DVT diagnosis and has markedly decreased the need for phlebography. Phlebography carries appreciable local morbidity, the risk of contrast administration, and is technically inadequate in 7-20% of studies. It is appropriate for use when falsenegative duplex imaging results are suspected on clinical grounds, or when the duplex study is technically inadequate.

Diagnosis of Pulmonary Embolism
Clinical recognition of possible PE.
There is no definitive set of bedside diagnostic findings. Clinicians select patients for testing for PE based on a high index of suspicion and awareness of clinical findings of PE illustrated in Table 1. The clinical features in Table 1 are listed in approximate order of positive predictive value, within each category. However, specific test characteristics for each finding are not available. The clinical detection of PE is not amenable to checklist or rule-based diagnosis; it remains a patternrecognition task, requiring the skills of an experienced clinician. Clinicians less familiar with PE are encouraged to consult an expert when the question arises.

Testing for PE.
The initial basis of testing for PE is ventilation-perfusion (V/Q) scanning. From 30-70% of
patients will need no other test, and a normal scan effectively excludes PE (see Figure 1 and Table 5). For other than normal tests, V/Q scanning returns a probability statement as a result, that must be evaluated in conjunction with the clinical findings (see Table 5). For example, the positive predictive value (PPV) of an indeterminate probability V/Q scan is 16% for patients where PE is considered clinically unlikely (middle section of Table 5); is 66% for patients where PE is considered likely (bottom section of Table 5), and 28% among those considered clinically uncertain. The algorithm in Figure 1 illustrates the approach taken with various combinations of clinical suspicion and V/Q findings. In general the only V/Q interpretations permitting direct clinical decisions are normal (no treatment) and high-probability (treatment). Pulmonary angiography is widely considered the reference standard for the diagnosis of pulmonary embolism. Without a higher standard to appeal to, we cannot discuss specificity and sensitivity of pulmonary angiography, using commonly accepted definitions of these terms. Instead, the accuracy of pulmonary angiography is discussed in terms of interobserver variability in the reading of pulmonary angiograms obtained in the context of large multicenter trials. Studies demonstrate that the larger the embolus, the larger the interobserver agreement. For segmental and larger emboli, agreement exceeds 95%. For subsegmental emboli, agreement is considerably less.

Preference for LMWH in DVT.
A number of high-quality randomized controlled trials (RCTs) have compared the several preparations of LMWH to UFH in the treatment of DVT. As summarized in two recent meta-analyses, LMWH for venous thrombosis confers a much lower risk of major bleeding complication (absolute risk reduction approximately 2 per 100 patients treated; relative risk
reduction of 58-68%), lower risk of recurrent thromboembolic disease (RRR 53-68%), and lower risk of death (RRR 47%). The data for pulmonary embolism are limited; to date LMWH appears as good as UFH for that indication, but sufficient data for recommendation for pulmonary embolism do not yet exist. Most studies of LMWH compare fixed-dose or weightadjusted-dose LMWH given subcutaneously to APTTadjusted- dose UFH given intravenously. It is not necessary to monitor LMWH therapy, and there are no routine clinical tests for doing so.

Route of administration.
LMWH is normally administered subcutaneously. Full dose UFH can be administered either by continuous intravenous (IV) infusion or by intermittent subcutaneous (SQ) injection. However, analyses of multiple randomized trials suggest that SQ UFH is as effective as IV UFH in the treatment of DVT, provided that an initial IV bolus dose (5- 10,000 U) is given, large doses of heparin are administered (usually > 17,500 U SQ BID), and heparin therapy is
monitored closely. Pulmonary embolus is currently treated with IV UFH. UFH can be administered as continuous IV infusion,intermittent IV boluses, or SQ boluses. Continuous
infusion is more readily monitored and adjusted, and probably achieves therapeutic levels more rapidly; hence it is the standard in our institution. There is only a single small study of patient preferences, which found that most patients preferred SQ administration, but IV equipment was not portable in that study.

Monitoring therapy.
LMWH does not require monitoring for therapeutic effect, and does not prolong APTT or TCT at therapeutic levels as much as does standard UFH. The effectiveness of UFH therapy is usually monitored by the activated partial thromboplastin time (APTT) or the thrombin time (also referred to as the thrombin clotting time, or TCT). The APTT is readily available and relatively inexpensive. Several studies have shown that anticoagulation guided by nomograms is superior to individual physician-guided therapy, which varies significantly. Published nomograms have been based on the APTT. Table 2 is one such nomogram, in which initial heparin dose is based on patient weight (the best predictor of heparin requirements), and subsequent dose changes are based on the APTT. An APTT ratio of 1.5-2.5 x control is generally considered therapeutic. Unlike the APTT, the TCT does not require addition of a thromboplastin, and it is not affected by acute phase increases in plasma proteins or coagulation factors, such as factor VIII, that occur in acutely ill patients. The TCT also exhibits a more linear relationship to plasma heparin concentrations than the APTT, especially in the supratherapeutic range. Therefore in hospitals in which the TCT is performed, a reasonable approach for dosing heparin is to calculate bolus and initial maintenance doses based on patient weight (bolus 80 U/kg, maintenance 18 U/kg/hr), and to base subsequent adjustments in heparin dose on the TCT (which is reported both in seconds and heparin units), aiming for a therapeutic range of 0.2-0.4 heparin units. In patients whose baseline APTT is prolonged (e.g. due to lupus-type inhibitor), the TCT is preferred over the APTT for monitoring heparin therapy. The APTT or TCT is usually measured every 6 hours until stable anticoagulation is achieved, then each morning. In patients receiving SQ heparin every 12 hours, clotting times are measured 6 hours after injection. Daily platelet counts are recommended for patients receiving UFH due to the approximately 5% incidence of heparin-induced thrombocytopenia (HIT). A modest and clinically unimportant reduction in platelet counts is more common than HIT. HIT is a serious complication of heparin therapy that can cause arterial and venous thrombosis, and less often bleeding. It is caused by a heparin-dependent platelet antibody that leads to platelet aggregation. The diagnosis should be suspected in a patient who develops thrombosis on heparin or when there is a fall
in platelet count to <100,000 or a decline by ≥50% from baseline counts during heparin therapy, or the appearance of venous or arterial thrombi. Monitoring of platelet counts should begin after the 4th day of heparin therapy, but earlier if the patient has previously been exposed to heparin. If the syndrome is suspected, stop heparin at once and consult with a specialist for testing and treatment options.

Overlap of Heparin and Warfarin
Heparin and warfarin therapy should overlap during the acute management of venous thrombosis. Clinical trials suggest that heparin can be discontinued safely once the INR enters the therapeutic range (2-3) if the patient has received > 5 days of heparin therapy. Some recommend that heparin be continued until the INR has been in the therapeutic range for > 2 days, since the antithrombotic effect of warfarin may be delayed relative to its effect on the prothrombin time. However, clinical trials have not tested whether this approach offers greater protection against thrombosis than discontinuation of heparin as soon as the INR is therapeutic.

Warfarin Anticoagulation
Efficacy.
Warfarin and other vitamin K antagonists reduce he incidence of recurrence of thrombosis in patients with DVT and pulmonary embolism by 30 or more per 100 patients treated.
Administration and monitoring.
Warfarin should be started early, usually within the first 24 hours of heparin therapy after heparin is therapeutic. Initial warfarin dosing is typically 10 mg on the first day followed by 5 mg qd thereafter, with doses given in the evenings. A target INR of 2.0-3.0 is effective in preventing thrombus extension or recurrence and is associated with a relatively low risk of bleeding. Combined analysis of 7 studies reveals that 19 of 1283 patients (1.5%) with venous thromboembolism experienced major bleeding during a 3 month course of warfarin with target INR 2.0-3.0. This equates to a major bleeding risk of 6%/yr in this patient population. Some patients, such as those with venous thrombosis and antiphospholipid antibodies, may require more intense warfarin therapy (i.e. INR 3.0-4.0). However, this point is controversial, and adequate studies addressing this specific issue are lacking.
Duration.
The optimal duration of warfarin therapy after DVT or PE depends upon clinical circumstances (see Table 3). Natural history studies suggest that after a first DVT the risk of recurrent thrombosis (PE or DVT) is 17.5% at 2 years, 25% at 5 years, and 30% at 8 years. Patients with continuing risk factors for thrombosis, such as malignancy, cardiomyopathy, immobility, or hypercoagulable states, are at higher risk, while patients who experience thrombosis under transient circumstances (e.g. post-operatively) are at lower risk of recurrence. In general, patients with a first episode of venous thrombosis should receive 3-6 months of warfarin (Table 3). Some studies suggest that patients with transient risk factors for recurrent thrombosis (e.g. post-operative DVT) can be safely treated with as few as 4 weeks of warfarin. However, these studies involved small numbers of patients, and one required documentation of a normal non-invasive venous study prior to stopping warfarin at 4 weeks. Given the low risk of major bleeding during properly monitored warfarin therapy (particularly in patients with transient risk factors for thrombosis), we recommend at least 3 months of warfarin after confirmed venous thrombosis. Six months of warfarin therapy after a first episode of DVT results in a lower rate of recurrence than 6 weeks of therapy.

Patients with a second episode of venous thromboembolism have a significantly lower rate of recurrence if they receive warfarin indefinitely (2.6% risk during 4 years of followup) as opposed to 6 months (20.7% risk of recurrence). However, this exposes the patient to a higher risk of bleeding complications. Prospective clinical trials addressing the optimal duration of warfarin therapy in patients with a first episode of venous thrombosis and an irreversible risk factor considered to place the patient at high risk of recurrence (e.g. malignancy, identifiable thrombophilia such as factor V Leiden) are lacking. However, recent studies suggest that certain patients within this heterogeneous group are at high risk of recurrent DVT. For example, the risk of recurrent DVT in patients with a first DVT who carry the factor V Leiden mutation is approximately 25% at 1 year. Therefore, some patients with a first episode of DVT and an irreversible thrombotic risk factor should be considered for indefinite warfarin therapy. However, each patient must be considered individually, and determination of the duration of therapy depends upon consideration of bleeding and thrombotic risks.

Other considerations.
The frequency of monitoring warfarin anticoagulation has not been rigorously studied. The frequency needed varies with both the patient's clinical condition and the stability of the PT level achieved. Monitoring daily is necessary at initiation, and weekly or more often during the first few weeks of therapy. Patients on long-established doses may be monitored as seldom as monthly. Patients on warfarin therapy should also be aware of the effect of both diet and drug interactions on their anticoagulation status. Information on dietary sources of vitamin K which can reduce the effect of warfarin should be provided as part of patient education, as should warning about OTC vitamin supplementation. Since the list of medications which interact with warfarin is lengthy, anticoagulated patients should be advised to ask their physician’s advice before taking any prescription or OTC medications, and be given a written list of potential interactions (such as a package insert or patient education sheet). Home monitoring of PT is being tested at the present time. Though not yet sufficiently advanced for this guideline to recommend it now, it may be addressed in future revisions.