Show
Deep vein thrombosis (DVT) is a common but elusive illness that can result in suffering and death if not recognized and treated effectively. DVT occurs in ≈2 million Americans each year. Death can occur when the venous thrombi break off and form pulmonary emboli, which pass to and obstruct the arteries of the lungs. DVT and pulmonary embolism (PE) most often complicate the course of sick, hospitalized patients but may also affect ambulatory and otherwise healthy persons.1234 It is estimated that each year 600 000 patients develop PE and that 60 000 die of this complication.567 This number exceeds the number of American women who die each year from breast cancer. PE is now the most frequent cause of death associated with childbirth.8 Women are a prime target for PE, being affected more often than men. Deep vein thrombosis is a major complication in orthopedic surgical patients and patients with cancer and other chronic illnesses. DVT can be a chronic disease. Patients who survive the initial episode of DVT are prone to chronic swelling of the leg and pain because the valves in the veins can be damaged by the thrombotic process, leading to venous hypertension. In some instances skin ulceration and impaired mobility prevent patients from leading normal, active lives. In addition, patients with DVT are prone to recurrent episodes. In those instances in which DVT and PE develop as complications of a surgical or medical illness, in addition to the mortality risk, hospitalization is prolonged and healthcare costs are increased. Over the past 20 years results of clinical trials have provided information that has revolutionized the approach to management of venous thromboembolic disease. New diagnostic modalities and therapeutic agents have been developed that are more effective, less expensive, and more convenient. Patients with venous thromboembolic disease (VTE) are seen by a variety of medical specialists, including general physicians, surgeons, obstetricians, hematologists, radiologists, and chest physicians. Because thromboembolic disease forms only a small part of the practice of most of these clinicians, it is difficult for them to keep abreast of advances that are important for optimal patient care. The purpose of this report is to provide medical trainees and clinicians with the information required to manage venous thromboembolic problems that they are likely to encounter in daily practice. Venous thrombi are intravascular deposits composed of fibrin and red cells with a variable platelet and leukocyte component.9 They usually form in regions of slow or disturbed flow in large venous sinuses and in valve cusp pockets in the deep veins of the calf (Fig 1) or in venous segments that have been exposed to direct trauma.101112 Venous thrombi often break off to form PE. The formation, growth, and dissolution of venous thrombi and PE reflect a balance between the effects of thrombogenic stimuli and a variety of protective mechanisms.131415 The factors traditionally implicated in the pathogenesis of venous thrombosis are activation of blood coagulation, venous stasis, and vascular injury.131415 Vascular damage contributes to the genesis of venous thrombosis through either direct trauma9121315 or activation of endothelial cells by cytokines (interleukin-1 and tumor necrosis factor) released as a result of tissue injury and inflammation. Blood coagulation can be activated by intravascular stimuli released at a remote site (eg, products of injured or infarcted tissue) or it can be activated locally by vessel wall damage (eg, damage to the femoral vein during hip surgery) or by cytokine-induced nondenuding endothelial stimulation.1215161718 These cytokines stimulate endothelial cells to synthesize tissue factor and plasminogen activator inhibitor-1 and lead to a reduction in thrombomodulin, thereby reversing the protective properties of normal endothelium. The thrombogenic effects of activation of blood coagulation are amplified by stasis and counteracted by rapid flow. Venous stasis predisposes the patient to local thrombosis by impairing the clearance of activated coagulation factors and limiting the accessibility of thrombin formed in veins to endothelial protein thrombomodulin, which is present in greatest density in the capillaries. The mechanisms that protect against thrombosis are inactivation of activated coagulation factors by circulating inhibitors, dilution and clearance of activated coagulation factors by flowing blood, inhibition of the coagulant activity of thrombin by thrombomodulin, enhancement of the anticoagulant activity of thrombin by thrombomodulin through activation of protein C, and dissolution of fibrin by the fibrinolytic system.19202122232425 Venous thrombosis in the lower limb can involve the superficial leg veins, the deep veins of the calf (calf vein thrombosis), the more proximal veins, including popliteal veins, the superficial femoral, common femoral, and iliac veins. Less commonly, thrombosis involves other veins in the body. Thrombosis of the superficial veins of the legs usually occurs in varicosities and is benign and self-limiting. Occasionally, however, the thrombi in superficial veins extend into the deep veins and give rise to major PE. Deep calf vein thrombosis is a less serious disorder than proximal vein thrombosis because thrombi in calf veins are generally small and are therefore not usually associated with clinical disability or major complications. Most calf vein thrombi are asymptomatic,10 but these thrombi can extend proximally and become dangerous. Venous thrombi produce symptoms because they obstruct venous outflow, cause inflammation of the vein wall or perivascular tissue, or embolize into the pulmonary circulation. Extension of thrombosis is more likely if the original thrombogenic stimulus persists. Complete spontaneous lysis of large venous thrombi is uncommon, and even when patients with venous thrombosis are treated with heparin, complete lysis occurs in fewer than 10% of cases.26 In contrast, complete dissolution of small, asymptomatic calf vein thrombi occurs quite frequently.10 There is a strong association between DVT and PE. Pulmonary emboli are detected by perfusion lung scanning in ≈50% of patients with documented DVT,327282930 and asymptomatic venous thrombosis is found in ≈70% of patients with confirmed clinically symptomatic PE. If the thrombus that embolizes is small (which is frequently the case when it is located in the calf), the embolus is usually asymptomatic and clinically insignificant, although the cumulative effect, if there are repeated showers of small emboli, can cause cor pulmonale. If the thrombus is large and involves the proximal veins, it often produces clinical manifestations; if it is very large or if the patient has a compromised cardiorespiratory system, it can be fatal. Most clinically significant and virtually all fatal emboli arise from thrombi in the proximal veins.1 Venous thrombi usually organize slowly and can be complicated by the postthrombotic syndrome.31 The residual abnormality can also act as a nidus for recurrent thrombosis,32 which occurs in approximately one third of patients over an 8-year follow-up period.33 Studies done before the introduction of anticoagulant therapy reported that the mortality rate for PE was ≈20% in hospitalized patients with clinically obvious venous thrombosis.34 In a small study, Kakkar and colleagues10 reported that without treatment, ≈20% of silent calf vein thrombi extended into the popliteal vein and that extension was associated with a 40% to 50% risk of clinically detectable PE. In a study of patients with clinically suspected DVT, Huisman and associates35 reported that 6.5% (20 of 307) who had negative impedance plethysmography at presentation developed evidence of extension over the next 10 days. Others have reported a lower frequency of impedance plethysmography (IPG) conversion during serial testing. The estimated frequency of extension rate of untreated symptomatic calf vein thrombosis is ≈30%, based on the results of these serial IPG studies. In contrast to untreated thrombosis, the short-term prognosis of patients with proximal DVT treated with adequate doses of anticoagulants for 3 months is good.363738 Clinically significant recurrent events take place in ≈5% of patients with proximal vein thrombosis treated with an initial course of heparin followed by oral anticoagulants or intermediate doses of subcutaneous heparin for 3 months.373839404142 Thereafter, DVT recurs in 5% to 10% of patients the year after anticoagulant therapy is discontinued363738 and in ≈30% of patients after 8 years.33 A comprehensive prospective follow-up study examining long-term prognosis in consecutive patients with a first episode of documented symptomatic DVT of the leg was recently completed by Prandoni and associates.33 The study assessed the long-term incidence of recurrent venous thromboembolism and postthrombotic syndrome. Patients were treated with an initial course of high dose–adjusted intravenous standard heparin or low-molecular-weight heparin (LMWH) followed by oral anticoagulants, which were started during the first week of treatment and continued for at least 3 months.42 The dose of oral anticoagulant therapy was adjusted daily to maintain the International Normalized Ratio (INR) between 2.0 and 3.0. All patients were instructed to wear graduated compression stockings (40 mm Hg at the ankle) for at least 2 years. They were seen at 3 and 6 months after presentation and every 6 months thereafter for follow-up assessments. Patients were asked to return immediately if they developed symptoms suggestive of recurrent venous thromboembolism. Follow-up continued for up to 8 years. A total of 355 consecutive patients with a first episode of DVT confirmed by venography were included in the study. Seventy-eight patients experienced one or more episodes of objectively confirmed recurrent venous thromboembolic events. Of the first recurrences, 35 (44.9%) occurred in a leg that was initially involved, 28 (35.9%) in the contralateral leg, and 15 (19.2%) were PE, which was fatal in 9 patients (11.5%). The cumulative incidence of recurrent VTE after 3 months was 4.9%; after 6 months it was 8.6%. The incidence of recurrent events gradually increased to 17.5% after 2 years, 24.6% after 5 years, and 30.3% after 8 years of follow-up (Fig 2). The risk of recurrent VTE was increased by the presence of malignancy and coagulation abnormalities and reduced in patients who had a reversible risk factor (eg, surgery and trauma or fracture). Of the 355 patients, 83 developed postthrombotic syndrome and 24 developed severe postthrombotic manifestations. The cumulative incidence of postthrombotic syndrome was 17.3% after 1 year and 22.8% after 2 years. Thereafter, the incidence of postthrombotic syndrome rose very gradually to 28.0% after 5 years and 29.1% at 8 years. Thus, in more than 80% of patients manifestations of postthrombotic syndrome became apparent in the first 2 years after acute thrombosis. The cumulative incidence of severe postthrombotic manifestations increased gradually from 2.6% after 1 year to 9.3% after 5 years. Thereafter, the cumulative incidence of severe postthrombotic manifestations did not increase further. It is likely that the use of compression stockings contributed to this low incidence of postthrombotic syndrome, as indicated by a recent controlled study.43 Ipsilateral recurrent DVT was associated with a strong increase in risk for postthrombotic syndrome (risk ratio 6:4). Surprisingly, there were no significant associations between occurrence of postthrombotic syndrome and size or location of the thrombus. Twenty-six of the 297 patients without a malignancy at baseline developed cancer. This occurred mainly in patients with idiopathic DVT at presentation.44 Of the 355 patients, 90 died during follow-up. The causes of death included malignancy (n=52), ischemic stroke (n=8), acute myocardial infarction (n=4), PE (n=9), heart failure (n=3), anticoagulant-related hemorrhage (n=2), and miscellaneous (n=6). In 6 patients who died suddenly, a definite cause of death was not established. Other studies have also reported that most recurrences take place in patients who have idiopathic venous thrombosis or who are exposed to a continuing risk factor (such as cancer). In these groups, the rate of recurrence is ≈15% in the 12 months after treatment is stopped. In contrast, the long-term prognosis in patients who develop venous thrombosis following exposure to a predisposing cause such as surgery or trauma is very good.45 Thus, provided they are treated with anticoagulants for 3 months,363738 fewer than 4% of these patients develop recurrences in the following year.454647 The label of recurrent venous thrombosis carries important prognostic implications. Patients are usually treated with anticoagulants for life and may suffer considerable mental anguish. Therefore, it is important to ensure that the diagnosis of recurrent DVT is correct. In many patients with clinically suspected recurrence, the diagnosis of recurrence is not confirmed by objective tests. For example, in a prospective study of patients with clinically suspected acute DVT, almost two thirds did not have this diagnosis confirmed by objective tests, and these patients did very well without anticoagulant therapy.48 The diagnosis of recurrent venous thrombosis can be difficult because venography, the diagnostic standard for acute venous thrombosis, is less reliable for diagnosis of recurrent venous thrombosis.48 However, the accuracy of diagnosis of acute recurrence has been improved by the introduction of noninvasive techniques (see below). In early descriptive studies, postthrombotic syndrome was reported to occur in ≈50% of patients with symptomatic venous thrombosis. More recently and possibly as a consequence of better initial anticoagulation and the use of graduated compression stockings, the incidence of postthrombotic syndrome after 8 years of follow-up was reported to be no more than ≈25%.33 The postthrombotic syndrome is caused by venous hypertension, which occurs as a consequence of recanalization of major venous thrombi leading to patent but scarred and incompetent valves or, less frequently, persistent outflow obstruction produced by large proximal vein thrombi.31495051 Recanalization and valve destruction result in a malfunction of the muscular pump mechanism, which leads to increased pressure in the deep veins of the calf. This high pressure results in progressive incompetence of the valves of the perforating veins of the calf, and when this occurs, flow is directed from the deep vein into the superficial system during muscle contraction, leading to edema and impaired viability of subcutaneous tissues and, in its most severe form, ulceration of venous origin. Follow-up studies of patients with proximal vein thrombosis have demonstrated that outflow obstruction (measured by IPG) is relieved either by recanalization or collateral flow in 30% of patients at 3 weeks and in 70% of patients at 3 months.52 Valvular incompetence is a more important cause of postthrombotic syndrome than is outflow obstruction.53 In patients with extensive thrombosis in the iliofemoral veins, swelling may never disappear, while in patients with less severe proximal vein thrombosis, swelling may subside after the initial event but return in the next few years. Other manifestations of postthrombotic syndrome are pain in the calf relieved by rest and elevation of the leg, pigmentation and induration around the ankle and the lower third of the leg, and, less commonly, ulceration and venous claudication, a bursting calf pain that occurs during exercise. Patients with extensive thrombosis involving the iliofemoral vein have a higher frequency of venous claudication and frequently have greater disability than patients with more distal vein thrombosis.50 However, incompetence of perforating veins may follow thrombosis confined to calf veins and may lead to stasis changes. In a follow-up study of calf vein thrombosis in Sweden, the frequency of postthrombotic syndrome was reported to be 13 of 79 or 16% in 2 years’ follow-up.54 There is evidence from recent studies that recurrent venous thrombosis is an important risk factor for development of postthrombotic syndrome33 and that risk of developing postthrombotic syndrome is reduced by the use of graduated compression stockings.43 The role of thrombolytic therapy in prevention of postthrombotic syndrome is uncertain. Clinical trials in acute DVT evaluating the effect of thrombolytic therapy on subsequent development of postthrombotic syndrome have produced equivocal results,55 although on balance, it is probable that the incidence of clinical symptoms is reduced in patients who receive thrombolysis.55 The prevalence of postthrombotic syndrome in the general population has been estimated in several countries. In Sweden it has been reported to occur in 2% of the population, and in a study of more than 4000 chemical-industry workers in Switzerland, the frequency of severe venous insufficiency with venous ulceration was reported to be between 1% and 1.5%.5456 In an investigation in Michigan involving more than 9000 adults older than 20 years, the prevalence of active or healed venous ulcers was 5 per 1000.2 Extrapolation of this figure to the general population in the United States suggests that about 500 000 Americans have or have had venous ulceration. The diagnosis of postthrombotic syndrome is sometimes obvious on clinical grounds if the symptoms are gradual in onset. However, patients can have subacute symptoms of leg pain and swelling, which may mimic acute recurrence of DVT. Although these symptoms are usually superimposed on a background of chronic pain and swelling, it may be difficult to exclude acute recurrence on clinical grounds alone, and a diagnosis of postthrombotic syndrome as the cause of the patient’s symptoms can be made only after acute recurrent venous thrombosis has been excluded. The diagnosis of postthrombotic syndrome should include demonstration of deep venous incompetence using Doppler ultrasound or plethysmography575859 and more recently by techniques such as volume plethysmography and duplex ultrasound. In some patients with recurrent leg pain not due to acute recurrent venous thrombosis or postthrombotic syndrome, an alternative cause is not found, and symptoms may be due to thromboneurosis. This clinical syndrome tends to occur in patients who have a morbid fear of the complications of DVT/PE. These patients may have had a previous episode of DVT and some have evidence of postthrombotic syndrome, but some have never had objectively documented episodes of venous thrombosis. These patients usually present with pain and tenderness that may be disproportionate to physical signs of swelling. In its most severe form, patients may be incapacitated by fear of recurrence, loss of the leg, or death. Patients frequently have a history of multiple hospital admissions for treatment of alleged recurrent venous thrombosis. Many are on long-term anticoagulant therapy or antiplatelet drugs, and some have undergone caval interruption procedures. Unfortunately, thromboneurosis is often iatrogenic, and fear of recurrence is reinforced each time the attending physician admits the patient to the hospital and orders treatment based on clinical suspicion alone. Thromboneurosis is best prevented by ensuring that a clinical suspicion of acute venous thrombosis (either first episode or recurrence) is always confirmed by appropriate objective tests. The most effective way of reducing death from PE and morbidity from postthrombotic syndrome is to institute a comprehensive institutional policy of primary prophylaxis in patients at risk for VTE. Patients can be classified as being at low, moderate, or high risk for developing VTE on the basis of well-defined clinical criteria60 (Tables 1 and 2), and the choice of prophylaxis should be tailored to the patient’s risk. In the absence of prophylaxis, the frequency of postoperative fatal PE ranges from 0.1% to 0.8% in patients undergoing elective general surgery, 0.3% to 1.7% in patients undergoing elective hip surgery, and 4% to 7% in patients undergoing emergency hip surgery.60 Safe and effective forms of prophylaxis are available for patients at high risk, and primary prophylaxis is cost-effective.61 Prophylaxis is achieved by either modulating activation of blood coagulation or preventing venous stasis. The following prophylactic approaches are of proven value: low-dose subcutaneous heparin,6162 intermittent pneumatic compression of the legs,6061 oral anticoagulants,6061 adjusted doses of subcutaneous heparin,63 graduated compression stockings,64 and LMWHs65 (Table 3). Antiplatelet agents such as aspirin are less effective for preventing VTE.60 Low-dose heparin is given subcutaneously at a dose of 5000 U 2 hours before surgery and is then given postoperatively at a dose of 5000 U every 8 or 12 hours. Low-dose heparin prophylaxis is the method of choice for moderate-risk general surgical and medical patients.60 Low-dose heparin reduces the risk of VTE by 50% to 70%62 ; it does not require laboratory monitoring and is simple, inexpensive, convenient, and safe. However, because of the potential for minor bleeding, it should not be used in patients undergoing cerebral, ocular, or spinal surgery. Low-dose heparin is less effective than warfarin,60 adjusted-dose heparin,63 and LMWH in patients undergoing major orthopedic surgical procedures.6566 Intermittent pneumatic compression of the legs enhances blood flow in the deep veins and increases blood fibrinolytic activity.60 This method of prophylaxis is free of clinically important side effects and is particularly useful in patients with a high risk of serious bleeding. Therefore, it is the method of choice for preventing venous thrombosis in patients undergoing neurosurgery,64 is effective in patients undergoing major knee surgery,67 and is as effective as low-dose heparin in patients undergoing abdominal surgery.60 Graduated compression stockings reduce venous stasis and are effective for preventing postoperative venous thrombosis in general surgical patients60 and in medical or surgical patients with neurological disorders, including paralysis of the lower limbs.64 In surgical patients the combination of graduated compression stockings and low-dose heparin is significantly more effective than low-dose heparin alone.6869 Graduated compression stockings are relatively inexpensive and should be considered for all high-risk surgical patients, even if other forms of prophylaxis are used. Moderate-dose warfarin (INR, 2.0) is effective for preventing postoperative VTE in all risk categories.60 Warfarin can be started preoperatively, at the time of operation, or in the early postoperative period. Although the full, measurable anticoagulant effect is not achieved until the third or fourth postoperative day, when treatment is started at the time of surgery or in the early postoperative period, warfarin is still effective in very high–risk patient groups, including patients with hip fractures.70 Prophylaxis with warfarin is less convenient than low-dose heparin or LMWHs because of the need for careful laboratory monitoring. Adjusted-dose heparin is given subcutaneously in a dose of 3500 U three times daily, starting 2 days before surgery. The dose is then adjusted to maintain the activated partial thromboplastin time (aPTT) at the upper limit of the normal range. Adjusted-dose heparin is more effective than fixed low-dose heparin in patients undergoing elective hip surgery63 but is less effective in preventing proximal vein thrombosis than LMWH following elective hip surgery.71 Adjusted-dose heparin is inconvenient because it requires careful laboratory monitoring. LMWHs have recently been approved for use as prophylactic agents in North America. LMWHs are safe and effective for prophylaxis in the following high-risk areas65 : elective hip surgery, hip fracture, major general surgery, major knee surgery, spinal injury, and stroke. LMWH has been reported to be more effective than standard low-dose heparin in general surgical patients,65 patients undergoing elective hip surgery,6566 and patients with stroke65 or spinal injury.65 In addition, LMWHs have also been more effective than warfarin in patients undergoing hip66 or major knee surgery,6566676869707172 and better than adjusted-dose heparin at preventing proximal vein thrombosis after elective hip surgery.71 Patients at moderate risk should be given prophylaxis (Table 3) with low-dose heparin. If anticoagulants are contraindicated because of an unusually high risk of bleeding, intermittent pneumatic compression should be used. LMWH, oral anticoagulants, or adjusted-dose heparin is effective following hip surgery. Of these three approaches, LMWH is the most convenient because laboratory monitoring is not required. Both LMWHs and intermittent pneumatic compression are effective in preventing venous thrombosis in patients undergoing major knee surgery. LMWH is more convenient and is the prophylactic method of choice. Intermittent pneumatic compression, with or without static graduated compression stockings, is effective and does not increase the risk of bleeding. A clinical suspicion of venous thrombosis should always be confirmed by objective tests because patients with minimal leg symptoms may have extensive venous thrombosis, whereas the classic symptoms and signs of pain, tenderness, and swelling of the leg can be caused by nonthrombotic disorders.117273747576 In most contemporary studies of ambulatory patients with symptoms compatible with venous thrombosis, the diagnosis of venous thrombosis is confirmed in only approximately one third when reliable objective tests are performed.767778 Alternative diagnoses include superficial thrombophlebitis, cellulitis, ruptured muscle or tendon, muscle strain, internal derangement of the knee, ruptured popliteal cyst, cutaneous vasculitis, and lymphedema.79 Despite the nonspecificity of clinical features, history and physical examination are important components of the diagnostic process because they may uncover an alternative cause of the patient’s symptoms and because they allow patients to be classified as having a high, intermediate, or low probability for venous thrombosis.80 With a simple clinical scoring system that included three main components (symptoms and signs at presentation, presence or absence of risk factors, and presence or absence of a possible alternative diagnosis), Wells and associates80 showed that ≈80% of patients with high clinical probability have venous thrombosis, while only 5% of patients with low clinical probability have venous thrombosis. When combined with the results of noninvasive tests, these pretest probabilities can be used to both simplify and reduce costs of the diagnostic process (Table 4). Although a number of tests have been evaluated over the years, only three have been shown to be accurate for diagnosing venous thrombosis in symptomatic patients: venography,818283 IPG,34357784858687888990 and venous ultrasonography.77919293949596979899100101102103104105 If used properly, any one of these methods is acceptable, although venous ultrasonography (also known as B-mode imaging) is the diagnostic method of choice in most patients with clinically suspected venous thrombosis.7799 In addition, the Simpli-red d-dimer test, which is performed on blood obtained by finger prick at the patient’s side and which has high sensitivity and moderate specificity, shows considerable promise as a test to rule out venous thrombosis.106 The d-dimer test is often false-positive after surgery or trauma, thereby limiting its value in these clinical situations. Venography is performed by injecting radiographic material into a superficial vein on the dorsum of the foot. The contrast material mixes with the blood and flows proximally. An x-ray image of the leg and pelvis will show the calf and thigh veins, which drain into the external iliac vein. With good technique, the entire deep venous system of the leg, including the external iliac and common iliac veins, may be imaged. A thrombus is diagnosed by the presence of an intraluminal filling defect.818283 Impedance plethysmography is performed by placing two sets of electrodes around the patient’s calf and an oversized blood pressure cuff around the thigh. The electrodes sense a change in blood volume (increased blood volume decreases electrical impedance) in the calf veins, which is recorded on a strip chart. Changes in venous filling are produced by inflating the thigh cuff to obstruct venous return and then reestablishing blood flow by deflating the cuff and assessing the time taken for venous volume in the calf to return to baseline. If an occlusive thrombus is present in the popliteal or more proximal veins, venous emptying is delayed. The test may also detect extensive calf vein thrombosis if venous outflow is obstructed, but it fails to detect the majority of calf vein thrombi.848687 Venous ultrasound imaging of the venous system is obtained with high-resolution equipment to produce two-dimensional images by real-time computation of reflected signals from an array of ultrasound sources.77949599102 The ultrasound probe is first placed over the common femoral vein in the groin. The transducer is then moved distally to visualize the superficial femoral vein over its course. The entire popliteal vein is then visualized in the popliteal fossa and traced distally to its trifurcation with the deep veins of the calf. Gentle pressure is applied with the probe to determine whether the vein under examination is compressible. The most accurate ultrasonic criterion for diagnosing venous thrombosis is noncompressibility of the venous lumen under gentle probe pressure.7799 Vein compressibility is best evaluated in the transverse plane. Visualization of the proximal portion of calf veins can often be achieved by experienced operators,95 but resolution can be suboptimal, and the sensitivity and specificity of venous ultrasonography is much lower for calf vein thrombosis than for proximal vein thrombosis. Duplex ultrasound, which combines real-time imaging with pulsed gated Doppler and color-coded Doppler technology, facilitates identification of veins, and as technology improves, diagnostic accuracy for calf vein thrombosis may increase.919293103104105106107 Although it has been claimed that color-coded Doppler is accurate for calf vein thrombosis, this contention has not been demonstrated by an appropriately designed clinical study. Venography is the reference standard, but it is invasive; the other two tests are noninvasive. All three tests are sensitive and specific for proximal vein thrombosis (thrombi in the popliteal and more proximal veins) in symptomatic patients, although IPG is less sensitive and less specific than venous ultrasound.108109110 Venography detects calf vein thrombosis. Venous ultrasonography detects ≈50% of symptomatic calf vein thrombosis; sensitivity is said to be higher in the hands of some experts, but this impression awaits confirmation in large clinical trials. Impedance plethysmography is insensitive to calf vein thrombosis, detecting <20%. Venous ultrasonography is now the diagnostic method of choice in patients with symptoms suggestive of DVT. Venography can be painful, it is relatively expensive and inconvenient to perform, and, on rare occasions, can be complicated by phlebitis. In addition, when performed by nonexpert radiologists, up to 30% of venograms are technically inadequate and therefore impossible to interpret. In contrast, venous ultrasonography is readily available, painless, and can be performed at bedside. However, like venography, this test is operator dependent. There is evidence from diagnostic studies using serial noninvasive testing in patients with symptoms of DVT that calf vein thrombi are not dangerous, provided that they remain confined to calf veins.33585111 However, calf vein thrombi can extend and do so in ≈30% of cases.74 Because only ≈5% of patients with symptoms of DVT have calf vein thrombosis (Fig 3),78 it is safe to exclude clinically important venous thrombosis if the venous ultrasonography is negative at presentation in patients who have low pretest clinical probability, because the negative predictive value of a negative venous ultrasound is more than 99%.80 In patients at moderate or high clinical probability, however, it would be prudent to repeat the test once after 5 to 7 days to detect the small percentage of patients with calf vein thrombosis that extends (Fig 4). The safety of withholding treatment when either the IPG or venous ultrasound test result is negative at presentation and subsequently on repeated testing over the next week has been demonstrated in a number of well-designed studies.33585111 Between 1% and 2% of patients with negative IPG at presentation and <1% of patients with negative venous ultrasonography develop clinically important events during the first 7 days of serial testing. When these patients with negative venous ultrasonography (or IPG) are followed up after 6 months, 99% have had no recurrences (Fig 5).111112 Diagnosis of Recurrent Venous ThrombosisThe diagnosis of clinically suspected recurrent venous thrombosis is often more difficult to establish than diagnosis of the first episode of venous thrombosis.48113 As with patients with suspected acute venous thrombosis, most patients referred with a diagnosis of recurrence do not have recurrent venous thrombosis. The clinical diagnosis of recurrent venous thrombosis is less specific than the diagnosis of the first episode of venous thrombosis48 because patients fear recurrence and physicians are sensitized to the possibility of this diagnosis. As a consequence, there is a tendency to overdiagnose recurrent venous thrombosis by attributing any new episodes of leg pain or swelling to a recurrent episode. Any other cause of leg pain or swelling can be confused with recurrence, but the most important mimic is postthrombotic syndrome, particularly because this disorder occurs in ≈30% of patients who have experienced proximal vein thrombosis.43114115116 The most common manifestations of postthrombotic syndrome, chronic aching and swelling of the calf, are unlikely to be confused with recurrent venous thrombosis. However, subacute exacerbations of pain and swelling can occur after episodes of increased activity or sometimes without an obvious precipitating cause and can be difficult to differentiate from recurrence. Because of their fear of recurrent venous thrombosis, patients often become concerned if they develop even minimal exacerbations of symptoms or signs. Finally, some patients develop recurrent episodes of superficial phlebitis or local cellulitis, which can be confused with recurrent DVT. For these reasons, and because overdiagnosis of recurrent venous thrombosis often results in unnecessary prolongation of anticoagulant treatment, every effort should be made to confirm a diagnosis of suspected recurrence. The diagnosis of recurrent venous thrombosis is made or excluded by a combination of either IPG and venography113 or venous ultrasonography and venography (Fig 6). A correct diagnosis of recurrent venous thrombosis is made by repeating the test used to make the initial diagnosis when the patient presents with suspected recurrence. The diagnostic process is facilitated by obtaining a baseline noninvasive test (either IPG or venous ultrasonography) when anticoagulants are discontinued and repeating the test if it is still abnormal at this time.48113 The negative test result can then be used as a baseline against which future tests can be compared.113 The rate of conversion is different for IPG and venous ultrasonography. The IPG result is negative in 60% of patients with proximal vein thrombosis by 3 months and in 90% by 12 months.51113 The rates of conversion for venous ultrasonography are lower than those for IPG.112117118 When the results of IPG or venous ultrasound are negative before presentation with a suspected recurrence, a positive result can be used to make a diagnosis of recurrent venous thrombosis. If the IPG performed at the previous visit was abnormal and remains abnormal at presentation with suspected recurrence, further testing with venography is required; if there is a new intraluminal filling defect, a diagnosis of recurrence can be made. If the results of venous ultrasound were abnormal at the previous visit, it is often possible to diagnose recurrence by demonstrating extension into a previously normal venous segment or by an increase in diameter of the venous lumen in a previously affected segment.112 Recurrence can be excluded if venography shows either no change or improvement compared with the previous examination or if a negative IPG or venous ultrasound remains negative on serial testing over the next 7 days (Fig 6). Diagnosis of Pulmonary EmbolismThe clinical diagnosis of PE is also highly nonspecific because the clinical features may be simulated by other cardiorespiratory or musculoskeletal disorders.3119120121122123124125126 Accordingly, the diagnosis should always be confirmed by objective tests. Patients may present with clinical features of minor or major PE. Patients with minor PE can have one or a combination of the following symptoms: transient shortness of breath, sharp localized chest pain aggravated by inspiration (pleuritic-type pain), and hemoptysis. The clinical features of minor PE are nonspecific and can also occur in patients with viral or bacterial pulmonary infections, postoperative atelectasis and pneumonia, acute bronchitis, and musculoskeletal chest wall pain. Esophageal spasm can cause severe chest pain that is not usually aggravated by breathing but may be confused with PE. Pleuritic-type chest pain may accompany pericarditis or immune pleuritis. In addition, patients with a past history of VTE may suffer anxiety attacks that are manifested as shortness of breath and occasionally as chest pain. These patients often have fleeting attacks of sharp chest pain that last for seconds or a feeling that they cannot take a deep breath. Patients with chronic obstructive lung disease who become acutely short of breath or develop pleuritic-type chest pain or hemoptysis present a difficult problem, because all of these complications can be produced by chest infection as well as by PE. Likewise, it can be difficult to differentiate between postoperative PE and postoperative atelectasis and infection, because both of these disorders can cause shortness of breath and pleuritic-type chest pain. Patients with major PE usually have severe shortness of breath with or without associated right-heart failure. Patients who sustain a massive embolism or have impaired cardiorespiratory reserve and sustain a moderate-sized embolus may present with hypotension, syncope, and peripheral circulatory failure. Sometimes there is associated dull central chest pain. Some of these features also occur in patients with acute myocardial infarction, a fulminating pneumonia, dissecting aortic aneurysm, pericardial tamponade, a massive hidden bleed, or septic shock. PE may also present with nonspecific manifestations such as arrhythmia, fever, unexplained heart failure, mental confusion, or, rarely, as bronchospasm. Approach to Diagnosis of Pulmonary EmbolismThe most reliable test for diagnosis of PE is pulmonary angiography, because a normal well-performed pulmonary angiogram excludes the diagnosis of PE, whereas demonstration of a constant intraluminal filling defect in a pulmonary artery establishes diagnosis.127128129130131132 However, pulmonary angiography is expensive, invasive, and not readily available in most hospitals and unavailable in many. Therefore, other less direct approaches are usually taken. The most useful test is the perfusion lung scan, because if the test result is normal, diagnosis of PE is excluded.133134 However, before the scan is performed, the patient should have a thorough clinical evaluation, because the combination of clinical probability and pulmonary scanning is important in clinical decision making. Using clinical features, presence or absence of risk factors, and presence or absence of features that suggest an alternative diagnosis, it is possible to classify patients into three groups: high, low, and intermediate probability. In addition, the patient should undergo chest radiography and electrocardiography. Although the latter tests are often not helpful,135136 they can be useful in ruling out other disorders that simulate PE. In addition, a chest radiograph is required for proper interpretation of the perfusion lung scan.137138 A second approach, which is complementary to the first, is to look for a source of PE in the deep veins of the leg with either venous ultrasound or venography. This approach can be very helpful, because although <20% of patients with proven PE have clinical symptoms or signs suggestive of leg vein thrombosis,3 ≈70% have venographic evidence of venous thrombosis.3 The perfusion scan remains the pivotal test. If the perfusion scan is normal, the diagnosis of PE is excluded. If the perfusion scan is abnormal, then the diagnostic approach depends on the clinical probabilities and the size and V/Q pattern of the defect. A diagnosis of PE can be made if the lung scan shows a segmental or greater perfusion defect and normal ventilation and the clinical probability is high or intermediate. A decision can be made to exclude a diagnosis of PE if clinical probability is low and the perfusion defect is small, particularly if it is matched (low-probability defect) (Table 5). All other combinations of clinical and lung scan probabilities require further investigation before a diagnosis of PE can be ruled in or out. In such patients, venous ultrasonography or venography is useful because a positive result allows a diagnosis of VTE to be made. Unfortunately, a negative test result for venous thrombosis cannot be used to rule out a diagnosis of PE because tests for venous thrombosis are negative in ≈30% of patients with established PE. The venogram or venous ultrasound may be negative for venous thrombosis in these patients because the source thrombus has embolized completely or because it originated in the deep femoral, internal iliac, or renal veins or the inferior vena cava, which are not usually visualized by venography. Alternatively, the embolism could have originated in upper limb veins, the right side of the heart, or the pulmonary arteries. Electrocardiography and Chest RadiographyWith PE, the ECG is often normal or shows nonspecific changes.135136 In patients with pericarditis or acute myocardial infarction, ECG changes may be diagnostic. In the appropriate setting, ECG changes of acute right-heart strain strongly suggest PE. The chest radiograph is rarely, if ever, diagnostic.119135 It may show a pneumothorax, pulmonary edema, or findings suggestive of primary or secondary malignancy. The finding of a Hampton hump (a semicircular opacity with the base abutting the pleural surface) is strongly suggestive of pulmonary infarction, but in the vast majority of patients chest radiography findings are nonspecific or normal. Other radiographic features compatible with PE include pleural effusion, subsegmental atelectasis, pulmonary infiltrate, raised hemidiaphragm, regions of apparent oligemia, or a prominent pulmonary vascular shadow at the hilum. However, none of these features are diagnostic of PE because they can be produced by other conditions, including obstructive lung disease, pulmonary infection, or atelectasis. Arterial Blood GasesMeasurement of arterial blood gases in patients with PE is rarely useful because arterial blood gas measurements lack specificity and are only moderately sensitive for PE.121125135 Hypoxemia and hypocarbia occur in conditions that simulate PE, and arterial oxygen tensions can be normal in patients with minor PE. Significant hypoxemia excludes hyperventilation as the cause of the patient’s symptoms, although this condition is rare. Lung ScansPerfusion scanning is performed by injecting isotopically labeled human macroaggregates of albumin intravenously. The macroaggregates are trapped in the pulmonary capillary bed and their distribution, which reflects the distribution of lung blood flow, is recorded with an external photoscanner. The perfusion lung scan is an important test because it is safe, readily available, essentially noninvasive, and, if entirely normal, rules out a diagnosis of PE.126133134 Ventilation scanning is performed with the use of radioactive aerosols that are inhaled and exhaled by the patient while a gamma camera records the distribution of the radioactivity in the alveolar spaces. An abnormal perfusion lung scan by itself is nonspecific and seen in a variety of cardiorespiratory disorders.3120122125137138 By combining perfusion and ventilation scanning, certain patterns occur that can be used to assign probabilities of PE.3122123137138139140 In general, the probability of PE is reflected in the size and pattern of perfusion defects. Thus, large defects are more likely to be caused by PE than small defects, and mismatched defects (abnormal perfusion and normal ventilation) are more likely to be caused by PE than are matched defects.3122123137138139140 However, these distinctions are not absolute. Thus, between 30% and 40% of patients with large perfusion defects with a matching ventilation defect have PE, and a small mismatched defect may not be diagnostic of PE.3122123 Patients with subsegmental perfusion mismatches have a probability of PE of ≈40%, and those with subsegmental matches have a probability of ≈25%.3122123 The probability is lower in patients in whom clinical suspicion of PE is low.3122123 A high clinical probability of PE combined with a high-probability lung scan pattern is associated with PE in 96% of patients.123 A moderate clinical probability combined with a high-probability lung scan pattern is positively associated with PE in 80% to 88% of cases.3123 In most circumstances, the presence of these combinations of clinical probabilities and lung scan findings can be used to make a clinical decision to diagnose PE and treat the patient accordingly. Unfortunately, these two combinations of clinical/lung scan patterns (ie, a high-probability lung scan with a high or moderate clinical probability) occur in only 12% to 32% of patients with abnormal perfusion scans.3123 In addition, only ≈50% of patients with a high-probability lung scan but a low clinical probability have PE.123 Although this combination is uncommon, it is important, because it would be inappropriate to make a diagnosis of PE without further investigation in this group. If both clinical probability and lung scan probability are low, then PE is very unlikely (occurring in <6% of patients), and for practical purposes a diagnosis of PE can be excluded.123 This combination of clinical/lung scan pattern occurs in ≈15% of patients with an abnormal lung scan. Thus, a management decision to either treat or not treat without further investigation can be made in <50% of patients with clinically suspected PE with an abnormal lung scan.3123 In the remaining patients with suspected PE and an abnormal perfusion scan, further investigations for venous thrombosis or PE are required to either rule in or rule out a diagnosis of PE.3123 An approach to diagnosis of venous thrombosis is shown in Fig 7. Approach to TreatmentThe objectives of treating venous thrombosis and PE are to prevent local extension of the thrombus, prevent the thrombus from embolizing, and, in certain clinical circumstances, accelerate fibrinolysis. Anticoagulants are effective in most patients for preventing clinically important local extension of thrombosis, but they must be continued for weeks to months after the acute event and they may not prevent long-term complications of thrombosis. Of the two anticoagulants in current use, heparin acts immediately by catalyzing the inhibition of activated coagulation factors (principally thrombin and factor Xa) by antithrombin III (AT-III), while coumarins act much more slowly by inhibiting synthesis of fully gamma-carboxylated vitamin K–dependent coagulation proteins. Both classes of anticoagulants inhibit the generation of factor Xa and thrombin when administered in relatively low doses. Oral anticoagulants do not inhibit thrombin activity directly but modulate further thrombin generation by lowering functional coagulation factors that participate in positive feedback loops. Heparin can inhibit thrombin activity as well as further thrombin generation by modulating positive feedback loops. Low concentrations of heparin can inhibit the early stages of blood coagulation, but higher concentrations are needed to inhibit the much higher concentrations of thrombin that are generated if the coagulation process resists modulation. If fibrin is formed, even higher concentrations of heparin are required to modulate the procoagulant effects of clot-bound thrombin, which is site-protected from inhibition by heparin/AT-III and can provide an ongoing procoagulant stimulus in the vicinity of the clot.141 Some of the new anticoagulants, including hirudin and its fragments, are effective inhibitors of clot-bound thrombin and may therefore be more effective than heparin in neutralizing the procoagulant effects of the fibrin-bound thrombin.141 The fibrinolytic enzymes streptokinase, urokinase, and TPA accelerate the rate of dissolution of thrombi and emboli. Thrombolysis is more expensive than anticoagulant therapy and is associated with a higher risk of bleeding, so its use should be restricted to patients who are likely to benefit from it. Two types of patient groups have the potential to benefit from thrombolytic therapy: those with major PE and selected patients with major venous thrombosis. Surgical removal of the thrombus (venous thrombectomy) or the embolus (pulmonary embolectomy) is rarely indicated. In patients with venous thrombosis, PE can be prevented very effectively with anticoagulant therapy. Pulmonary emboli can also be prevented by inserting a filter into the vena cava, but this approach is used only if anticoagulant therapy is contraindicated because of bleeding or if PE has recurred despite adequate treatment with anticoagulants (see below for definition of adequate anticoagulant therapy). There is good evidence that patients with PE have a high mortality and a high rate of recurrence if untreated.142 There is also good evidence that patients with symptomatic proximal143144 or calf vein thrombosis145 have a high recurrence rate without treatment. Anticoagulation reduces mortality and recurrence in patients with acute PE and reduces recurrence in patients with DVT.143144145 Use of Anticoagulant TherapyA working approach to the use of anticoagulants is described below. A more comprehensive guide, “Guidelines to Anticoagulant Therapy,” (Circulation 1994;89:1449-1489) is available in reprints from the Office of Scientific Affairs, American Heart Association, 7272 Greenville Ave, Dallas, TX 75231-4596. (Telephone 800-242-8721.) HeparinHeparin should be initiated with an intravenous bolus of 5000 U followed either by an intravenous infusion of 1400 U/h or a subcutaneous injection of ≈17 500 U twice daily.146147148 A weight-adjusted dose regimen can also be used.149 This regimen consists of a continuous intravenous infusion in a bolus dose of 80 U/kg followed by an infusion at 18 U/kg per hour. The aPTT should be performed ≈6 hours after the bolus and initiation of the continuous infusion and at least daily thereafter to maintain the aPTT in the therapeutic range equivalent to an anti–factor Xa heparin level of 0.3 to 0.7 U/mL. Warfarin can be started within the first 24 hours. Heparin is continued for 5 days150151 or longer until prothombin time (PT) has been in the therapeutic range for a minimum of 2 consecutive days. It is essential that the initial dose of heparin be adequate to achieve a therapeutic aPTT and that the period of overlap of heparin and warfarin is sufficient to allow the full antithrombotic effects of warfarin to be expressed (Table 6). The distinction between expression of the anticoagulant and antithrombotic effects of warfarin is discussed in a subsequent section of this report. Therapeutic RangeThe concept of a therapeutic range is based on experimental studies in animals152 and subgroup analysis of the results of two prospective studies in humans.41153 The animal studies demonstrated that prevention of growth of experimental venous thrombi required doses of heparin that prolonged the aPTT to approximately twice that of control subjects. These doses were equivalent to a heparin level of 0.2 U/mL by protamine titration of the thrombin time. In the clinical studies, comparisons of the rates of recurrence between patient subgroups demonstrated that risk of recurrence was increased if the aPTT ratio was less than 1.5 times the mean of the normal range. The results of these studies have led to the recommendation that the therapeutic range of heparin should be an aPTT ex vivo (ie, measured on plasma of patients treated with heparin), which is equivalent to a heparin level by protamine titration of the thrombin time of 0.2 to 0.4 U/mL or an anti–factor Xa heparin level of 0.3 to 0.7 U/mL. For many commercial aPTT reagents, the therapeutic range is ≈1.8 to 3.0,98 although for less sensitive reagents it is 1.5 to 2.0154 (Table 7). A large between-patient variation in dosage is required to achieve a therapeutic aPTT response in patients with VTE.146155 With a continuous infusion of heparin started at a dose of 32 000 U per 24 hours after a bolus of 5000 U, approximately one third of patients are below the therapeutic range at 6 hours, one third are in the therapeutic range, and one third are above the therapeutic range.146 By adjusting the dose according to a specially developed dose-adjustment nomogram146 (Table 8) in which the aPTT response is obtained every 6 hours until the therapeutic range has been achieved, more than 80% of patients are within the therapeutic range at 24 hours and more than 90% are within this range at 48 hours.146 The anticoagulant effect of heparin is influenced by its nonspecific binding to plasma proteins that compete with AT-III for heparin binding and by the rate of heparin clearance.156157 Many of the heparin-binding proteins are acute-phase reactants that are elevated to a variable degree in sick patients. The elevated levels of these plasma proteins contribute to heparin resistance seen in sick patients, and the variable concentrations of these binding proteins contribute to the differences in anticoagulant response among patients.156157 One of these acute-phase reactant proteins, factor VIII, also reduces the effect of heparin on the aPTT; thus, in many sick patients the observed resistance to heparin and variability in dose response is greater when monitored with an aPTT than with an anti–factor Xa heparin assay or a thrombin clotting time.158 Differences in the rates of heparin clearance between patients also contribute to interindividual variability in patients’ responses. Treatment of patients who fail to achieve an adequate aPTT response despite high doses of heparin has been clarified by the results of a randomized trial.158 Patients with venous thrombosis whose aPTT response to high doses of heparin (more than 35 000 U per 24 hours) was subtherapeutic were randomly allocated to monitoring with either aPTT or a heparin level of 0.3 to 0.7 anti–factor Xa units. These therapeutic ranges (for both methods of monitoring) correspond to a heparin level by thrombin time protamine titration of 0.2 to 0.4 U/mL.159 Many patients who had subtherapeutic aPTT values had heparin levels >0.3 U/mL. In patients randomly assigned to monitoring by aPTT, the dose of heparin was increased until the test result was in the therapeutic range. Despite receiving a lower dose of heparin, patients randomly assigned to monitoring by heparin level had a low rate of recurrence that was no different than the group randomly assigned to monitoring with aPTT. Heparin assays based on an anti–factor Xa assay or a thrombin time should be used to monitor heparin therapy in patients who have a long aPTT due to a “lupus anticoagulant.” The targeted therapeutic range should be 0.3 to 0.7 anti–factor Xa units or 0.2 to 0.4 U/mL by protamine titration of the thrombin time.159 Heparin AssaysHeparin assays using a chromogenic substrate are easy to perform in any clinical laboratory, although they are not often available clinically. Heparin assays are more expensive than aPTT assays; therefore, it is recommended that their use be limited to the 10% to 20% of patients whose aPTT response is below the lower limit of the therapeutic range with heparin doses of 40 000 U per 24 hours. In patients with an inadequate response to heparin therapy by both the aPTT and heparin assay, the dosage of heparin is increased, and an assay for AT-III is obtained. If the AT-III level is <50% of normal, the patient is treated with infusions of plasma or AT-III concentrate to elevate the AT-III level. However, if the AT-III level is above 60%, the dose of heparin is increased with the use of a heparin dose-adjustment nomogram.146 Duration of Heparin TherapyThe practice of a 7- to 10-day course of heparin therapy has been changed because of the findings of two randomized studies performed in patients with DVT. The studies reported that a 4- to 5-day course of heparin was as effective as a 9- to 10-day course of heparin.150151 The results of these two studies have important practical implications because the shorter course of heparin facilitates early discharge of patients from the hospital. Although the findings of these studies can likely be generalized to most patients, they may not be applicable to patients with large iliofemoral vein thrombosis or major PE, because these two classes of patients were excluded from one study150 and formed only a small proportion of patients in the second.151 It is our practice to treat patients with large iliofemoral vein thrombi and those with major PE with a 7- to 10-day course of heparin and to delay starting warfarin therapy until the aPTT has been in the therapeutic range for 3 days. The delay in starting warfarin is used to ensure that patients receive an adequate dose of heparin for at least 5 days. Subcutaneous HeparinThe relative efficacy and safety of heparin administered by subcutaneous and continuous intravenous infusion have been compared in randomized trials. These studies demonstrate that the two methods are equally safe and effective, provided that heparin is given in an adequate starting dose and that the dose is adjusted according to the aPTT (Table 9).148160161162163 However, there is a clinically important reduction in the bioavailability of heparin when administered subcutaneously in doses up to 15 000 U twice daily that results in subtherapeutic anticoagulant and antithrombotic effects in a large percentage of patients.153 On the other hand, there is good evidence from one large study that heparin administered subcutaneously is both safe and effective when started at a dose of 17 500 U twice daily after an intravenous bolus of 5000 U. The dose is then adjusted according to the aPTT.148 The aPTT is performed 6 hours after the morning injection and the dose is then adjusted to maintain the midinterval aPTT at 1.5 to 3.0 times the control value.148 Dose estimation is a little more difficult than with continuous infusion, but the feasibility of this approach has been demonstrated in a number of clinical trials.148160 Subcutaneous administration is difficult in patients in shock or heart failure because of poor and variable subcutaneous tissue blood flow. Low-Molecular-Weight HeparinsAdministration of LMWHs in a fixed dose by subcutaneous injection has been compared with administration of dose-adjusted heparin by continuous infusion for treatment of venous thrombosis. The results, which have been summarized in a meta-analysis,164 indicate that LMWHs are at least as effective and safe as standard heparin. These findings raise the possibility that selected patients with venous thrombosis might be suitable candidates for treatment at home, an advance that would reduce cost and improve patient convenience. Like heparin,165 LMWHs do not cross the placental barrier,166167168 and descriptive studies suggest they might be safe and effective171 in pregnancy. In a randomized trial LMWHs were associated with a much lower incidence of heparin-induced thrombocytopenia than heparin170 and a lower incidence of osteoporosis.164 Oral AnticoagulantsThe need for oral anticoagulants after an initial course of heparin is based on the results of two randomized studies that demonstrated that the incidence of out-of-hospital recurrences could be markedly reduced if heparin therapy was followed by a 3-month course of warfarin.143145 In one study in which the dose of warfarin was adjusted to obtain an INR of 3.0 to 4.5, the incidence of bleeding was very high. Another study was then conducted in which patients with proximal vein thrombosis were randomly assigned to treatment with either high- (INR, 3.0 to 4.5) or moderate-intensity (INR, 2.0 to 3.0) warfarin after an initial course of heparin therapy.37 The incidence of recurrence was equally low in both groups, but bleeding was approximately four times higher in the high-intensity group. Based on the results of this study, and subsequent experience with other prospective clinical studies, the recommended therapeutic range is an INR of 2.0 to 3.0. An INR of 3.0 to 4.0 has been recommended for patients with antiphospholipid antibodies,171172173 although there is some disagreement on this issue.174 Antithrombotic Effect of WarfarinWarfarin therapy is usually monitored by prothrombin time (PT), a test that is responsive to reduction of 3 of the 4 vitamin K–dependent procoagulant clotting factors (factors II, VII, and X). The conventional view is that the antithrombotic effect of warfarin is reflected by its anticoagulant effect as measured by PT. However, this view may not be correct during the induction phase of warfarin therapy. During the first few days of warfarin therapy, PT primarily reflects the reduction of factor VII activity, which has a half-life of only ≈6 hours, which is similar to the half-life of the natural anticoagulant protein C. Subsequently PT is prolonged by depression of factors X and II (prothrombin). Therefore, for the first 24 hours of warfarin therapy there is potential for a transient hypercoagulable state, resulting from a reduction of levels of protein C before the effects of warfarin on the activities of factors X and II are fully expressed. There is evidence that reductions of factor II and, possibly, factor X are more important than reduction of factors VII and IX for the antithrombotic effect of warfarin. The evidence supporting this view comes from the following observations. First, the experiments of Wessler and Gitel,175 performed more than 40 years ago with a stasis model of thrombosis in rabbits, showed that the antithrombotic effects of warfarin require 6 days of treatment, whereas the anticoagulant effect of warfarin as reflected by prolongation of PT is seen within 2 days. These findings are consistent with an explanation that the antithrombotic effect of warfarin requires a reduction in activity of factor II, which has a half-life of ≈60 hours. Second, in more recent experiments in a rabbit model of tissue factor–induced intravascular coagulation, Zivelin et al176 demonstrated that the protective effect of warfarin primarily reflects its ability to lower factor II levels. Thus, selective infusion of factor II, and to a lesser extent factor X, abolished the protective effects of warfarin in this model. In contrast, infusion of factor VII or IX had no effect. The concept that the antithrombotic effect of warfarin reflects its ability to lower factor II levels provides a rationale for overlapping heparin with warfarin in treatment of patients with thrombotic disease until the factor II level is lowered into the therapeutic range. Given that factor II has a half-life of ≈60 hours, an overlap of at least 4 days is necessary. Optimal DurationPatients with VTE are usually treated with oral anticoagulants for 3 to 6 months. Shorter courses of oral anticoagulant therapy have been investigated in randomized trials, but the results have been inconclusive.177178179 It is now clear that risk of recurrence varies in different subgroups. The risk of PE in patients with isolated calf DVT is very low.85 There also is evidence46180 that risk of recurrence is less in patients with a temporary or reversible risk factor (eg, thrombosis secondary to surgery or trauma) than it is in those with a continuing risk factor (such as associated malignancy) or with idiopathic DVT (thrombosis in the absence of a recognized risk factor). Prandoni and associates180 reported that in patients with proximal vein thrombosis treated with oral anticoagulants for 3 months, the rate of recurrent VTE was 24% over 80 weeks in patients with idiopathic venous thrombosis compared with 4.8% in those with a reversible risk factor. A similar observation was made by Levine and associates.46 In 301 patients with proximal DVT given 3 months of warfarin and then followed for an additional 9 months, there were 26 recurrent thromboembolic events in 212 patients (12.3%) with either continuing risk factors or idiopathic DVT, compared with 0 in 89 patients with a transient reversible risk factor (P=.0007). None of the recurrences were fatal. Thus, patients with an identifiable reversible risk factor (such as surgery) appear to respond well to a 6-week to 3-month course of therapy, whereas patients without a reversible risk factor have a high incidence of recurrence despite 3 months of oral anticoagulation. Similar findings have been reported in two randomized studies. In the first report, 712 patients with DVT and PE were randomly assigned to either 4 or 12 weeks of anticoagulant therapy.47 The rate of recurrent VTE was 7.8% in patients treated for 4 weeks and 4.0% in those treated for 12 weeks. Only 1 of 116 patients (0.86%) with postoperative VTE had a recurrent event; whereas among the 506 “medical” patients, 4.0% of patients treated for 12 weeks and 9.1% of patients treated for 4 weeks experienced a recurrence. Only 1% of all patients had fatal PE. These results suggest that a short course of anticoagulation might be adequate for patients with postoperative thrombosis, but a longer course of treatment is necessary for patients without a reversible risk factor. In a more recent study,181 897 patients with a first episode of DVT or PE were treated with at least 5 days of heparin or LMWH and randomly allocated to receive 6 weeks or 6 months of warfarin, with the goal of reaching an INR of 2.0 to 2.85. The incidence of recurrence over 2 years of follow-up was 18.1% in the 443 patients who received 6 weeks of oral anticoagulation compared with 9.5% in the 454 patients who received 6 months of therapy (P<.001). However, as in the other studies, the incidence of recurrent thromboembolism was much lower in both groups in patients with reversible risk factors. The observed difference in recurrence rates between patients with and without reversible risk factors is relevant to the issue of optimal duration of oral anticoagulant therapy. Thus, the low absolute incidence of thrombosis in patients with temporary risk factors suggests that a short course of treatment might be appropriate for the subgroup of patients with reversible risk factors, whereas long-term anticoagulant therapy should be considered for patients without a reversible predisposing factor. At present, however, there is insufficient evidence to support lifelong treatment for all patients with idiopathic thrombosis. Instead, it would be reasonable to use anticoagulant therapy for 6 weeks in patients with a reversible risk factor and to continue anticoagulation for up to 6 months in patients with idiopathic venous thrombosis. An indefinite duration of anticoagulation should be considered in patients with venous thrombosis associated with active malignant disease who are often bedridden and receiving chemotherapy, which contributes to their hypercoagulable state.182 Long-term anticoagulant therapy should also be considered for patients who have multiple recurrent episodes of idiopathic VTE and those with inherited thrombophilia who have suffered one or more unprovoked episodes of major VTE.45 Recommendations for Duration of Warfarin TherapyPatients with a first episode of VTE should be treated for 6 weeks to 3 months if they have a reversible risk factor and for 3 to 6 months if they have idiopathic venous thrombosis. Warfarin therapy should be continued for longer periods, possibly for life, in patients with documented idiopathic thrombosis who have 1 of the 4 inherited molecular abnormalities (deficiencies of AT-III, protein C, protein S, or activated protein C resistance) and in those who have a lupus anticoagulant or anticardiolipin antibody, because these laboratory abnormalities predispose them to recurrent venous thrombosis. Treatment of patients with these blood abnormalities who develop venous thrombosis after a well-recognized provocation (eg, surgery) is uncertain. Indefinite anticoagulation might not be warranted, although some authorities believe so. The AHA also recommends that patients who have more than two documented episodes of recurrent venous thrombosis and patients with at least one episode of thrombosis and active cancer should be treated with anticoagulants indefinitely. Finally, patients with ongoing risk factors (eg, immobilization in a plaster cast) should be treated until the period of risk is over. Most patients requiring long-term anticoagulant therapy respond well to warfarin targeted to an INR of 2.0 to 3.0. However, some patients with cancer have a resistance to warfarin and require long-term treatment with heparin, administered in full doses by subcutaneous injection. The optimal intensity of anticoagulation therapy is uncertain for patients with a lupus anticoagulant or cardiolipin antibody who require long-term anticoagulation. There are reports, based on retrospective analyses of observational studies, that patients with the antiphospholipid antibody syndrome and thrombosis are inadequately protected from recurrent episodes of VTE if treated at a targeted INR of 2.0 to 3.0.171172173 In contrast, a recent smaller prospective study in lupus anticoagulant–positive patients with venous thrombosis174 but without other manifestations of the antiphospholipid antibody syndrome reported that these patients with fewer complications respond well to warfarin at an INR intensity of 2.0 to 3.0. It is uncertain whether the discrepant findings reported in these studies result from differences in patient populations or differences in the responsiveness of PT reagents to the lupus anticoagulant in patients who receive anticoagulation with warfarin. Thus, it is possible that with some PT reagents the INR result is artifactually prolonged by the lupus anticoagulant and therefore does not reflect the true anticoagulant effects of warfarin. Thrombolytic TherapyThrombolytic therapy is more effective than heparin in producing rapid lysis of thromboemboli. However, it is more expensive than heparin, it is associated with a higher risk of bleeding,55183 and it is not indicated in most patients with PE because they do well clinically with anticoagulant therapy. It is contraindicated in the postoperative period and in other situations in which there is a high risk of bleeding. Thrombolytic therapy has lifesaving potential for patients with massive PE184185 and should be considered in patients with major PE who have syncope, hypotension, severe hypoxemia, or heart failure.184185186 Thrombolytic therapy should also be considered for patients with a submassive embolism and underlying cardiac or respiratory disease. Limited evidence suggests that thrombolytic therapy prevents postthrombotic syndrome in some patients with acute venous thrombosis of recent onset.183187 Thrombolytic therapy may also be indicated in selected patients (both young and old without risk factors for bleeding) with extensive proximal vein thrombosis.187188 Caval InterruptionAlthough anticoagulation is the standard treatment for acute venous thrombosis and PE, venous interruption procedures may be indicated for VTE when anticoagulation is ineffective or unsafe. The most common indication for venous interruption in patients with DVT or PE is anticoagulant-induced bleeding or anticipation of hemorrhagic complications in a patient with a predisposing lesion, such as a bleeding peptic ulcer, gastrointestinal malignancy, recent intracranial operation, or an underlying hemorrhagic state (eg, liver failure or thrombocytopenia). The second indication for venous interruption is failure of anticoagulation, provided that the anticoagulant effect has been within the prescribed therapeutic range (an aPTT corresponding to an anti–factor Xa heparin level >0.3 U/mL or an INR >2.0). Development of new PE or substantial extension of venous thrombosis should be documented by objective tests before recurrent thromboembolism is accepted as a diagnosis, because new symptoms in a patient with an established venous thrombosis of PE are often misinterpreted as evidence for recurrence in a patient receiving anticoagulation. Other indications for venous interruption are more controversial. These include • Major PE with severe cardiovascular instability • As an adjunctive procedure in patients who undergo pulmonary embolectomy • Prophylactic interruption of the inferior vena cava in patients at exceptionally high risk of VTE, particularly if there is a relative contraindication to anticoagulation Intracaval DevicesThe first intracaval device to be widely used was the inferior vena caval umbrella devised by Mobin-Uddin et al.189 The umbrella filter is inserted through a cutdown in the internal jugular vein and passed under fluoroscopic control through the superior vena cava and right atrium into the inferior vena cava, where its position below the renal veins is confirmed by phlebography. When it is expelled from its capsule applicator, the pointed struts engage the wall of the cava and hold the filter in place. The device contains fenestrations to maintain venous blood flow. The results following implantation of 4699 filters during the first 6 years after the umbrella filter became available were summarized by Mobin-Uddin.190 The initial design had a diameter of 23 mm and was associated with proximal migration in 27 of 2848 applications (0.9%). The frequency of proximal migration was reduced to 0.4% by increasing the diameter to 28 mm. Complete occlusion of the filter occurred in 30% to 45% of patients due to thrombosis around the device or trapping of an embolus.190191 The reported rate of recurrent PE was 12%.192 Less common complications included perforation of adjacent organs (eg, duodenum or ureter) and breakage. The Greenfield filter has essentially replaced the Mobin-Uddin umbrella. The filter, which resembles an umbrella consisting only of struts, is placed with its apex directed proximally. With this design, emboli are retained in the center of the cone, where the spokes are closer together and the trapping efficiency greater.193 The central positioning of entrapped emboli facilitates blood flow past the trapped embolus and may encourage fibrinolysis, thereby accounting for the high rate of patency with this device (95%).191 Magnant and coworkers194 reviewed the experience with placement of the Greenfield filter. They concluded that percutaneous placement of inferior vena caval filters had supplanted operative placement and that no major morbidity had been associated with use of the Greenfield filter. The bird’s nest filter was invented by Roehm and described in 1984.195 When compared directly with the Greenfield device, the bird’s nest filter appeared to be more readily dislodged and more easily subjected to local thrombosis.195 Occlusion of the cava by a balloon has been proposed by Hunter et al196 and Moser et al.197 The balloon is inserted as a percutaneous procedure. A potential advantage of balloon occlusion is that caval obstruction can be temporary; once the threat of embolization has subsided, the balloon can be deflated and removed. However, thrombosis can occur around the balloon. In a report of up to 18 years of experience involving 191 cases, Hunter and associates198 reported no malfunction of the inflation mechanism and no migration from the site of inflation. No patients had recurrent PE after balloon inflation. In 39% of patients, the legs appeared normal and free of edema. Surgical RemovalThrombectomy for acute venous thrombosis and pulmonary embolectomy for acute PE to relieve acute obstruction are rarely used. Thrombectomy is of limited benefit because it is usually complicated by acute recurrence despite postoperative anticoagulant therapy; it leaves a de-endothelialized venous surface that is highly thrombogenic.199 However, it is indicated to rapidly reduce venous obstruction in patients with phlegmasia cerulea dolens with impending venous gangrene. Although pulmonary embolectomy can be a lifesaving procedure in a patient with massive embolism,200 most hospitals do not have the resources, personnel, or facilities for this type of surgery. Furthermore, most patients who are likely to benefit from pulmonary embolectomy die before they can be diagnosed and treated,201 and some candidates for emergency pulmonary embolectomy survive and do well with medical therapy. On the other hand, elective pulmonary thromboendarterectomy can be very effective and lifesaving in selected patients with chronic large-vessel thromboembolic pulmonary hypertension.202203 This operation, which has been available for years but abandoned by many centers because of a high postoperative mortality, has been revived by the work of the San Diego group.202203 The success of the procedure is highly dependent on the availability of a skilled and experienced team of surgeons and internists. Upper-Extremity DVTThe frequency of upper-extremity venous thrombosis involving the axillary and/or subclavian veins has increased in the last decade with the increasing use of long-term indwelling catheters. Upper-extremity venous thrombosis is classified as primary and secondary. Primary upper extremity thrombosis can be caused by local venous compression produced by unusual movements or positions of the arm (“effort thrombosis”), whereas secondary thrombosis is usually caused by indwelling intravenous devices.204205 Effort thrombosis has been described following weight-lifting, pole-vaulting, racquet sports, or direct and prolonged pressure to the axilla. Axillary vein thrombosis can also be a manifestation of the thoracic inlet syndrome or can be caused by direct trauma or compression by tumor.206 Axillary or subclavian vein thrombosis has been described in AT-III deficiency, protein S deficiency, hypoplasminogenemia, and antiphospholipid syndrome.207208209210 Thrombosis secondary to the use of long-term indwelling catheters, often used in administration of chemotherapy, is now a much more common cause of upper-extremity thrombosis than effort thrombosis. Upper-extremity venous thrombosis can be complicated by PE211 and rarely by massive PE.212213 The most important complications are long-term disability caused by venous hypertension and loss of venous access in patients requiring long-term chemotherapy. Venous hypertension can produce swelling, fatigability, aching, and weakness of the affected arm, particularly following activity. The symptoms can be disabling in athletes or manual laborers during and after activity involving the affected arm. The reported frequency of disabling upper-extremity venous hypertension after spontaneous axillary/subclavian vein thrombosis varies from 25% to 47%.214215216 Lower rates (12%) have been reported in a series of patients treated with thrombolysis,217 but no randomized trials have been reported comparing anticoagulants with thrombolysis. The diagnosis of upper-extremity vein thrombosis is usually suspected on clinical grounds and confirmed by venography. Optimal visualization of the thrombosed axillary/subclavian veins is best achieved by injecting the radiographic contrast into the median basilic vein. Injection of contrast material into a distal vein in the hand or wrist will demonstrate an obstruction and the presence of collateral vessels but does not usually outline the thrombus. Imaging studies in which color flow duplex ultrasound was used lack the sensitivity of venography for upper-extremity thrombosis.218 Various treatments have been advocated for primary upper-extremity thrombosis. Resolution of acute symptoms can usually be obtained with either anticoagulant or lytic agents.145146147 Anticoagulant therapy is not usually associated with anatomic resolution of the thrombus and clinical improvement because collaterals develop and bypass the obstruction. Thrombolytic therapy appears to be more effective than anticoagulants in producing early resolution.219220 Local therapy administered through a small catheter introduced through the basilic vein and advanced into the clot has been advocated. A loading dose of 250 000 IU urokinase infused into the clot over 1 hour and then continued at a lower dose of 1000 IU/min for up to 24 hours has been used successfully.221 The patient is then treated with heparin for 5 days, followed by warfarin for 3 months. Surgical removal of the first rib has been advocated by some if symptoms of venous obstruction persist after a course of conservative treatment. However, the effectiveness of this invasive approach has never been evaluated in an appropriately designed clinical trial. Long-term venous access through a central venous catheter is required for treatment of long-term disorders requiring chemotherapy, antibiotics, or hyperalimentation. Thrombosis of the subclavian/axillary vein is a common complication of central venous catheterization. These thrombi may be asymptomatic,222223 although spontaneous resolution is uncommon when long-term venographic follow-up studies are performed.224 The standard treatment of secondary axillary/subclavian vein thrombosis has been removal of the catheter, limb elevation, and anticoagulation. This approach usually results in rapid improvement of symptoms, but on follow-up 70% of patients have been reported to have some pain and/or swelling in the affected arm.225 Of greater importance, the venous lumen is obliterated and cannot be used again for venous access. Thrombolytic therapy has been used successfully to treat secondary upper-extremity thrombosis.226227228 Initial reports used high-dose systemic therapy. More recently local catheter-directed thrombolytic therapy has been used with apparent success.229 A dosage regimen of urokinase has been established, empirically consisting of 250 000 IU/h for 2 hours followed by 60 000 IU/h until clot lysis has been achieved. Heparin can be given in full doses either during or after completion of thrombolytic therapy and anticoagulation with heparin, followed by warfarin for ≈3 months. With this approach, a 78% lysis rate has been reported in a small study of 31 patients.229 Successful lysis is more common with fresh thrombi.230 Diagnosis of Venous Thromboembolism in the Pregnant PatientThe diagnosis of VTE during pregnancy is difficult because leg pain and swelling are frequent and usually not due to DVT,231 and performance of radiological procedures is a problem because of the fear of exposing the fetus to radiation. Deep Vein ThrombosisAs in the nonpregnant patient, venous ultrasonography is used as the initial diagnostic test. If venographic confirmation of an equivocal test result is required, a limited venogram can be performed without risk to the fetus by covering the patient’s abdomen with a lead-lined apron. A limited venogram allows visualization of the calf veins, popliteal vein, and most of the superficial femoral vein but not the iliac vein. Therefore, a normal limited venogram does not exclude iliac vein thrombosis. Pulmonary EmbolismThe diagnosis of PE in pregnancy is essentially the same as in the nonpregnant patient, with three exceptions designed to avoid exposure of the fetus to ionizing radiation: (1) Ventilation and perfusion scanning are performed at 50% of the usual dose; (2) pulmonary angiography, if indicated, should be performed via the brachial route rather than the femoral route; and (3) venography, if indicated, should be limited, with shielding of the abdomen. Chest radiography, perfusion, and ventilation lung scanning are performed with a reduced dose of radioisotope for the perfusion scan (1 to 2 MCi). If the perfusion scan is normal, PE is excluded; if the lung scan indicates a high probability of PE, the diagnosis is made and the patient is treated with anticoagulants. If the scan is nondiagnostic, the patient is investigated for DVT by IPG or duplex ultrasound; if the test results are abnormal, the patient should be treated with anticoagulants. If the results are normal, a pulmonary angiogram should be considered. Management of Venous Thromboembolism During PregnancyAn excellent review of this subject has been published,232 and the recommendations outlined below follow this report, which should be read for additional details. HeparinA recent critical review of the literature of heparin therapy during pregnancy233 reported that, contrary to a previous report,234 heparin therapy during pregnancy is safe for the fetus. The conclusion is corroborated by a cohort study in which the rates of premature birth, spontaneous abortion, stillbirth, neonatal death, and congenital malformation were not significantly higher in 100 pregnant women treated with heparin than in the normal population.235 Because heparin does not cross the placenta, there is no increased risk of bleeding for the fetus. WarfarinIn the review cited previously, the pooled rate of adverse effects associated with warfarin therapy was high (26.1%).233 Warfarin exposure between 6 and 12 weeks of gestation can be associated with warfarin embryopathy, which is characterized by stippled epiphyses and nasal hypoplasia. In a study by Iturbe-Alessio et al,236 10 of 35 term pregnancies in which warfarin was administered between 6 and 12 weeks were associated with warfarin embryopathy. This is likely to be an overestimate, and the true incidence of warfarin embryopathy is likely to be ≈5% of infants if maternal exposure occurs between 6 and 12 weeks of gestation. Warfarin embryopathy has not been reported with warfarin exposure outside this time period. Central nervous system abnormalities, both hemorrhage and malformations, have been reported after warfarin exposure at any time during pregnancy, but the incidence is very low.233236 Therefore, heparin is the anticoagulant of choice for treatment of VTE during pregnancy. If warfarin is used, it should be restricted to the second and early third trimesters and avoided between 6 and 12 weeks of gestation and near term to avoid delivery of an anticoagulated fetus. Treatment of Acute Deep Venous Thrombosis and Pulmonary EmbolismHeparin is usually initiated with an intravenous bolus of 5000 U followed by a maintenance dose administered as a continuous intravenous infusion of 32 000 U per 24 hours to prolong the aPTT into the therapeutic range (≈1.8 to 2.5 times control for most reagents) for 5 to 7 days. After the initial intravenous dose of heparin, subcutaneous heparin should be administered every 12 hours in doses adjusted to prolong a 6-hour postinjection aPTT into the therapeutic range. The aPTT should be checked regularly, because heparin requirements may vary as pregnancy progresses. The patient should be monitored three times in the first week and then at least weekly thereafter. Anticoagulant therapy should be continued throughout pregnancy and for 4 to 6 weeks after delivery. If the episode of VTE occurs late in pregnancy, anticoagulation should be continued for a total of 3 months after the episode. Long-term Anticoagulant Therapy Before PregnancyPatients who receive long-term warfarin therapy before pregnancy for DVT/PE or prevention of systemic embolism should be treated with subcutaneous heparin every 12 hours throughout pregnancy in doses adjusted to prolong the 6-hour postinjection aPTT to ≈1.5 to 2.5 times control. Two options are available when patients receiving long-term anticoagulant therapy decide to conceive. The first is to switch the patient to heparin before conception. This has the advantage of avoiding any exposure of the fetus to warfarin but increases the duration of heparin exposure if conception is delayed. The second option is to continue warfarin and perform regular pregnancy tests when conception is attempted. As soon as the pregnancy test result is positive, warfarin should be stopped and heparin started. This is probably safe for the fetus, provided warfarin is discontinued before 6 weeks of gestation. As mentioned above, no cases of fetal embryopathy have been described with warfarin exposure before 6 weeks of gestation. Warfarin therapy can be resumed after delivery. Previous Deep Venous Thrombosis and Pulmonary EmbolismThe optimal treatment of pregnant patients with previous DVT/PE is unknown because there are no large prospective trials to provide reliable estimates of the incidence of recurrence during pregnancy. Prophylaxis with standard heparin, 5000 U every 12 hours, is a reasonable approach and is associated with a very low recurrence rate.235 Surveillance with weekly IPG or duplex ultrasonography may be a reasonable alternative to heparin during pregnancy. Delivery and PostpartumIf the patient is receiving 5000 U of heparin every 12 hours at term, heparin can be discontinued at onset of labor. No increase in bleeding is anticipated with this approach. If adjusted-dose heparin is being administered at term, some pregnant patients can have a prolonged aPTT for as long as 20 hours after their last dose of subcutaneous heparin.237 To overcome the potential risk of a long aPTT at delivery, elective induction can be planned and heparin therapy discontinued 24 hours before induction. In patients considered to be at high risk for thrombotic complications, an intravenous heparin infusion can be started after discontinuation of subcutaneous heparin. Because the half-life of intravenous heparin is short,238 heparin can be discontinued 4 to 6 hours before delivery with the expectation that the aPTT will be normal at time of delivery. After delivery, heparin and warfarin should be restarted as soon as hemostasis is obtained, and heparin can be discontinued after an appropriate period of overlap. When administered to the nursing mother, warfarin is safe for the breastfed infant.239240 Other Therapeutic ModalitiesThere are very few reports on the use of thrombolysis during pregnancy. As a general rule, pregnancy is a relative contraindication to thrombolytic therapy, and its use should be restricted to patients with massive PE.241242 LMWHs do not cross the placenta and have been used successfully during pregnancy.169243244 Their advantage over standard heparin is a more predictable dose response and a longer half-life after subcutaneous injection, which allows administration once daily without frequent monitoring.244 Management of Venous Thromboembolism in ChildrenVenous thromboembolism in children is much less common than in adults. In general, recommendations for antithrombotic therapy have been extrapolated from those used for adults. However, optimal dosing for antithrombotic therapy in children is likely to differ from adults because the anticoagulant response to antithrombotic agents is different. IncidenceIncidence of DVT/PE in the adult population is ≈2.5% to 5.0%.254245 In comparison, incidence of DVT/PE in the general pediatric population is reported to be 0.07 per 10 000 and 5.3 per 10 000 hospital admissions.246247248 Other comparisons illustrating the lower risk of DVT/PE during childhood are the <1% incidence of clinically apparent DVT/PE after lower limb or scoliosis surgery249 and the relative absence of DVT/PE in children with congenital thrombophilias.250251 Clinical FeaturesNinety-five percent of DVT/PE in pediatric patients occurs as a complication of serious diseases such as prematurity, cancer, trauma/surgery, and congenital heart disease.250251252253 Congenital prethrombotic disorders account for <10% of DVT/PE in children.250251 Children at greatest risk for DVT/PE are younger than 1 year or teenaged.250251252253 DVT in the lower extremities is the most frequent noncentral venous line thrombotic complication in children.251 The clinical presentations of DVT and PE are similar to those in adults.248250251253 Central Venous LinesForty percent of DVT in children and more than 80% in newborns occurs in the upper venous system secondary to use of central venous lines,250251252 which are employed for short-term intensive care or long-term supportive care in children requiring total parenteral nutrition or therapy for cancer. Central venous line–related DVT requires repeat anesthesia for replacement and can be complicated by PE254255256257 ; it can cause superior vena cava syndrome257258259260261 and chylothorax257258262263 and can obliterate the upper venous system264265 and so lead to postthrombotic syndrome in the upper extremities. Treatment of ChildrenChildren older than 2 months who have DVT or PE should be treated with intravenous heparin (bolus 75 U/kg; initial maintenance of 20 U/kg per hour) sufficient to prolong the aPTT to a range that corresponds to an anti–factor Xa level of 0.3 to 0.7 U/mL. Treatment with heparin should be continued for 5 to 10 days and oral anticoagulation overlapped with heparin for 4 to 5 days. For many patients heparin and warfarin can be started together and heparin discontinued on day 6 if the INR is therapeutic. Heparin therapy should be used for a longer period for massive PE or iliofemoral thrombosis. Long-term anticoagulant therapy should be continued for at least 3 months, with oral agents (initial dose 0.2 mg/kg per day) to prolong PT to an INR of 2.0 to 3.0. Either indefinite warfarin therapy with an INR of 2.0 to 3.0, low-dose anticoagulant therapy (INR, <2.0), or close monitoring should be considered for children with a second recurrence of venous thrombosis or a continuing risk factor such as central venous line, antithrombin deficiency, or protein C or S deficiency. Newborns with a central venous line in place should be treated with intravenous heparin in doses of 1 to 5 U/h through the catheter.266267268269270 Treatment of NewbornsThe optimal regimen for anticoagulation therapy in treatment of newborns with DVT, PE, or arterial thrombosis is uncertain. If anticoagulation is indicated, a short course (10 to 14 days) of intravenous heparin (75 U/kg bolus; maintenance 28 U/kg per hour), sufficient to prolong the aPTT to an anti–factor Xa level of 0.3 U/mL, should be used. The role of thrombolytic agents in treatment of VTE is uncertain. Further clinical investigation is needed before more definitive recommendations can be made. If thrombolytic therapy is used, either urokinase or TPA is preferable to streptokinase, and supplementation with plasminogen may be helpful. Complications of AnticoagulationTreatment of patients who develop complications during anticoagulant therapy involves management of the actual complication and subsequent management of the thromboembolic event for which the patient is being treated. Bleeding is by far the most important complication of anticoagulant therapy. The approach to bleeding depends on the severity of bleeding, the anticoagulant and dose used, results of laboratory tests at the time of bleeding, the length of time the patient has been treated with anticoagulants, and the seriousness of the thromboembolic event for which the patient is being treated. HeparinThe frequency of clinically important bleeding during a 5- to 10-day course of heparin therapy varies between 3% and 10%, depending on whether the patient is at high or low risk.148151153160162271272 In many cases bleeding is not life-threatening and does not require discontinuation of heparin. Because heparin has a relatively short circulating half-life (60 minutes),273274275276 the anticoagulant effect is reversed fairly rapidly after treatment is discontinued. In most cases bleeding is treated by discontinuing heparin, applying local pressure as appropriate, and replacing blood if necessary. If bleeding is potentially life-threatening (eg, intracerebral, intraspinal, retroperitoneal, or severe gastrointestinal), heparin should be stopped and the anticoagulant effect reversed with protamine sulfate. Protamine sulfate is a strong basic substance that rapidly neutralizes the effect of heparin. The appropriate neutralizing dose depends on the dose of heparin and route and time of administration. If protamine sulfate is used within minutes of intravenous heparin injection, then a full neutralizing dose, 1 mg protamine per 100 U heparin, should be given. Since the half-life of heparin is ≈60 minutes, only 50% of a full neutralizing dose is required 1 hour after the last heparin injection, and only 25% of the full neutralizing dose is required after 2 hours.147 Protamine sulfate can produce a hypotensive response if given rapidly, so the dose should be injected slowly over a 20-minute period.277278279 Some patients may also develop a hypersensitivity reaction to protamine sulfate. Heparin rebound may occur if very large doses of heparin are given.280281282283 Therefore, it may be necessary to repeat administration of protamine if laboratory tests demonstrate a residual heparin effect.169 A direct assay of heparin activity, thrombin time, or aPTT should be performed both before and immediately after protamine is infused, and the test should be repeated 2 hours later to determine whether the neutralizing effect of protamine on heparin is permanent or transient. If bleeding occurs when the aPTT response is in the therapeutic range or just beyond the therapeutic range, or if the anticoagulant-associated bleeding is potentially life-threatening, treatment with anticoagulant therapy should be stopped, and an alternative form of treatment should be used to manage the thromboembolic event. If the patient has proximal vein thrombosis or major PE, a caval interruption procedure should be considered.284 If the patient has calf vein thrombosis, the course of the thrombus can be monitored with serial venous ultrasound imaging99111 and a caval interruption procedure used if thrombosis is extended. The risk of bleeding is influenced by five variables: the patient’s clinical condition,151 the dose of heparin,285286 the anticoagulant response,286287 method of administration,151288 and concomitant use of aspirin or thrombolytic agents.289290291292 The most important risk factor for bleeding is recent surgery or trauma. Other risk factors are renal failure, old age, and peptic ulcer disease. There is a relation between bleeding and both heparin dose and anticoagulant effect.285286287293294 Bleeding is greater when heparin is administered by intermittent intravenous injection.162288 Other Complications of Heparin TherapyOther complications of heparin are thrombocytopenia,295296 with or without thrombosis296 ; osteoporosis,297298299300301302 which occurs only with long-term treatment; and local skin hypersensitivity and skin necrosis confined to subcutaneous injection sites.303 Other complications are very rare and include anaphylaxis, hypoaldosteronism,304305306 and alopecia. In addition, patients treated with heparin can develop hyperkalemia307 and often develop an asymptomatic increase in plasma levels of hepatic transaminases.308 If a patient develops local skin reactions at the site of injection, the source of heparin should be changed because local reactions may not occur with a different preparation of heparin, including LMWHs. ThrombocytopeniaThrombocytopenia is a well-recognized complication of heparin therapy. Two forms of thrombocytopenia are described: an early benign, reversible nonimmune thrombocytopenia and a late, more serious IgG-mediated immune thrombocytopenia. The mechanism of the early form, which is not associated with adverse clinical sequelae, is uncertain but could be the result of direct weak activation of platelets by heparin.309310311312 The immune form of heparin-induced thrombocytopenia is characterized by strong IgG-mediated platelet activation170296313 and is associated with a substantial risk of thrombotic complications. The incidence of serologically confirmed heparin-induced thrombocytopenia was investigated in a large clinical trial that compared unfractionated heparin (7500 U twice daily) with LMWH (30 mg enoxaparin twice daily) for prophylaxis after elective hip surgery.170 The incidence of heparin-induced thrombocytopenia was ≈1% at 7 days and ≈3% at 14 days in patients receiving unfractionated heparin and 0% in those receiving LMWH. Other prospective studies with higher (therapeutic) doses of heparin have reported a similar incidence of thrombocytopenia.314315316317318319320321322323324 Heparin-induced thrombocytopenia usually begins between 5 and 15 days after the start of heparin therapy (median, 10 days),170295325 but it has been reported within hours of starting heparin in patients who have received heparin within the previous 3 months.295303326 Thrombosis associated with heparin-induced thrombocytopenia can be heralded by a fall in platelet count without overt thrombocytopenia (eg, from 350 000 to 150 000). For this reason, patients who receive heparin should undergo a platelet count daily, and if the platelet count falls by 50% or more, heparin should be stopped and an alternative management strategy instituted. Thrombocytopenia and Paradoxical ThrombosisHeparin-induced thrombocytopenia is a highly prothrombotic disorder. In a large prospective study of heparin therapy after elective hip surgery, risk for thrombosis was dramatically increased (odds ratio, 37) in patients with heparin-induced thrombocytopenia, compared with those who did not develop it.173 Although many case series have emphasized the association of heparin-induced thrombocytopenia with arterial thrombosis (“white clot syndrome”), it is now clear that venous thrombosis is much more common with heparin-induced thrombocytopenia than arterial thrombosis.170327 Overall, prospective studies suggest that thrombosis associated with heparin-induced thrombocytopenia occurs in ≈1% of patients who receive unfractionated heparin for more than 5 days.170324 Bleeding complications have been described in patients with heparin-induced thrombocytopenia, but they are less frequent and much less important than thrombotic complications.325 Laboratory Manifestations and PathogenesisTypically, the platelet count nadir in heparin-induced thrombocytopenia is between 20 and 150 000 per milliliter (median nadir, 50 000).325 Approximately 5% of patients have concomitant hypofibrinogenemia associated with disseminated intravascular coagulation.325 The platelet count usually returns to baseline levels within 1 week of discontinuing heparin. Heparin-induced thrombocytopenia is caused by an IgG that activates platelets via their FcκII receptors.328329 The major target antigen is a heparin sulphate/platelet factor IV complex that localizes the IgG on the platelet surface.330331332333 The thrombogenic diathesis results from in vivo platelet activation334 as well as generation of procoagulant platelet-derived microparticles.335 In addition, heparin-induced thrombocytopenia IgG has been shown to activate endothelium in vitro via recognition of a heparin sulfate/platelet factor IV complex.332333336 Laboratory TestingPlatelet activation assays that use washed target platelets337338 have a sensitivity and specificity for heparin-induced thrombocytopenia of at least 95%.170 Typically, heparin-induced thrombocytopenia IgG activates platelets at low (0.5 to 1.0 U/mL) but not high (10 to 100 U/mL) concentrations of heparin.337339340 Aggregation studies in which citrated plasma is used are much less sensitive to heparin-induced thrombocytopenia IgG than assays in which washed platelets are used.340341342 An ELISA assay with the platelet factor IV/heparin target antigen has been developed330 that shows good concordance with the platelet activation assay.342 Although heparin-induced thrombocytopenia is much less common with LMWH preparations than standard heparin, in vitro studies indicate that LMWHs show immune cross-reactivity in ≈70% of instances.170343 In contrast, in vitro cross-reactivity is much less common (≈10%) with the heparinoid Orgaran,344 which has been used successfully as a substitute for heparin in patients with heparin-induced thrombocytopenia.345 TreatmentTwo different antithrombotic agents have been evaluated in descriptive studies. These are Orgaran344345 and the defibrinogenating snake venom ancrod (Arvin).346347 Intravenous administration of Orgaran produces immediate onset of anticoagulation after bolus administration. Ancrod has the advantage of exhibiting no cross-reactivity with heparin, but there is a delay of ≈12 hours before effective defibrinogenation can be achieved. In addition, neither thrombin generation nor platelet activation are inhibited by ancrod348 and the magnitude of the anticoagulant effect is less predictable than with Orgaran. Ancrod is also contraindicated in patients with disseminated intravascular coagulation or septicemia.349 Long-term (ie, >3 weeks) anticoagulation with ancrod is limited by development of antibodies that render patients resistant to its effects.346350 Unfortunately, neither of these agents is approved for use in the United States, but they can be obtained for compassionate use. Initial experience with hirudin from Europe is very promising, but it is not approved for use in North America. Complications of Oral AnticoagulantsBleeding is by far the most common complication of oral anticoagulant therapy.351 Randomized studies have shown that the risk of bleeding is influenced by the intensity of anticoagulation,37352353354 and several studies have shown that the risk of clinically important bleeding is reduced by lowering the therapeutic range for the INR from 3.0 to 4.5 to 2.0 to 3.0. Although this difference in anticoagulant intensity is produced by a mean reduction in the dose of warfarin of only ≈1 mg, the effect on bleeding is profound. Randomized studies have also shown that the rate of oral anticoagulant–induced bleeding is increased by concomitant use of high doses of aspirin that both impair platelet function and produce gastric erosions.351355356 Multivariate analysis of cohort studies also suggests that risk of bleeding is influenced by the underlying clinical disorder.354357 These studies reported that the risk of major bleeding is increased by age >65 years, a history of stroke or gastrointestinal bleeding, and the presence of serious comorbid conditions such as renal insufficiency or anemia.358359360 Bleeding that occurs when the INR is <3.0 is frequently associated with an obvious underlying cause37 or an occult gastrointestinal or renal lesion.353 Drugs that are known to interact with coumarins should be avoided if possible.361 However, if concomitant use of drugs that interact with warfarin is necessary, PT should be monitored more frequently in the first few days to weeks of combined use to anticipate a change in dosage. Furthermore, all new drugs should be viewed as having the potential to interact with coumarins, and the frequency of PT monitoring should be increased in the initial period after introduction. Drugs known to inhibit platelet function should be avoided unless prescribed to augment the antithrombotic effects of warfarin. For example, low-dose aspirin (100 mg/d) augments the antithrombotic effects of coumarins in patients with prosthetic heart valves but at an increased risk of minor bleeding.362 The frequency of bleeding depends very much on intensity of the anticoagulant effect and patient-related risk factors.352358363 If moderate-dose anticoagulant therapy is used to prolong the INR to between 2.0 and 3.0, bleeding is relatively uncommon.352358363364365366367 Most episodes occur in patients with a potential bleeding source such as a peptic ulcer, gastritis, renal calculus, or malignancy. Bleeding complications in patients on long-term anticoagulant therapy tend to occur early and may unmask an underlying local source. In randomized trials of moderate-intensity warfarin (INR, 2.0 to 3.0) in patients with nonvalvular atrial fibrillation versus untreated control subjects, the typical annual incidence of major bleeding was between 1.0% and 1.5% in the warfarin groups and 0.5% to 1.0% in the control groups. However, patients selected for these trials were at low risk for bleeding, so in practice, bleeding on warfarin is higher than reported by these studies. Management of BleedingIf bleeding occurs during oral anticoagulant treatment in a patient with VTE, management depends on severity of bleeding, INR at the time of bleeding, and whether or not the patient has completed most of the prescribed course of anticoagulant therapy. If the INR is above the therapeutic range, treatment can be discontinued until bleeding has stopped and then reintroduced cautiously at a lower intensity. If the INR is within the therapeutic range, a local source of bleeding should be sought, particularly if bleeding is gastrointestinal or from the urinary tract. However, if the INR is markedly prolonged, it is not usually necessary to look for a source of bleeding. If bleeding is life-threatening and the INR prolonged, the coagulation defect should be reversed immediately by infusion of plasma, and vitamin K1 should be administered in a dose of 10 mg to 25 mg either intravenously by slow infusion or by subcutaneous injection. If bleeding is not life-threatening and the INR is markedly prolonged, then the anticoagulant effect can be reversed by administering 5 mg vitamin K1 by subcutaneous injection. Vitamin K1 can interfere with subsequent warfarin therapy when doses of 10 mg or more are used, and it can cause refractoriness to further warfarin therapy for up to 2 weeks. Skin NecrosisThe most important nonhemorrhagic side effect of warfarin is skin necrosis. This uncommon complication is usually observed on the third to eighth day of therapy353354357360365366368369370371372373 and is caused by extensive thrombosis of the venules and capillaries within the subcutaneous fat. An association has been reported between warfarin-induced skin necrosis and protein C deficiency,368372373 and less commonly, protein S deficiency,374 but this complication can also occur in persons without a deficiency. A role for protein C deficiency seems probable and is supported by the similarity of the lesions to those seen in neonatal purpura fulminans, which complicates homozygous protein C deficiency. The reason for the unusual localization of the lesions to subcutaneous fat deposits remains a mystery. The optimal technique for initiating anticoagulant therapy in patients with known protein C or protein S deficiency is uncertain. A reasonable empirical approach is to start with an initial course of heparin, begin warfarin at a maintenance dose of 5 mg, and give both anticoagulants in combination for ≈7 days. In patients who develop warfarin-induced skin necrosis, warfarin should be discontinued, vitamin K1 should be given to increase levels of protein C, and full doses of heparin should be administered to achieve a rapid anticoagulant effect. Treatment of patients with warfarin-induced skin necrosis who require anticoagulant therapy for an indefinite period is difficult. These patients can be treated with subcutaneous heparin long term, but this is inconvenient and carries a risk of osteoporosis. It might be safe to reintroduce warfarin in low doses initially in combination with heparin and to use combined treatment for 10 to 14 days, during which time the warfarin dose is gradually increased.373 It should be noted, however, that heparin may not terminate coumarin necrosis,375376 and some have reported that heparin failed to prevent continuing skin necrosis in homozygous protein C deficiency with very low protein C levels.377378379380 Management When Anticoagulants Are StoppedManagement of thromboembolism is influenced by the nature of the thromboembolic event, the time during the course of anticoagulant therapy that bleeding occurred, and the INR level during bleeding. If bleeding occurs in a patient with calf vein thrombosis who has received an adequate course of heparin therapy, then oral anticoagulant therapy can be stopped and replaced with low-dose heparin 5000 U twice daily SC. If bleeding occurs toward the end of a course of anticoagulant therapy (eg, >2 months after starting treatment) in a patient with proximal vein thrombosis, a decision can be made to terminate the course of anticoagulants. Long-term Warfarin Therapy and Elective SurgeryLong-term anticoagulation is indicated in patients with a sustained high risk of arterial or venous thromboembolism. Elective surgery in such patients can be difficult. A number of approaches are available, but none have been evaluated in appropriately designed clinical studies. In general, the management choice should take into account the risk of thromboembolism if the patient is left untreated and the risk of bleeding if aggressive perioperative-operative anticoagulation is used. The least complicated approach is to stop oral anticoagulants and perform elective surgery when the INR has returned to the normal range. Oral anticoagulants can then be started postoperatively in combination with low-dose or full-dose heparin, the choice of heparin regimens depending on the anticipated risk of postoperative bleeding. White and associates381 have reported that it takes ≈4 to 5 days for an INR between 2.0 and 3.0 to return to the normal range after warfarin is discontinued. Stopping anticoagulants 4 to 5 days preoperatively is appropriate in patients with atrial fibrillation or mechanical prosthetic valves because the risk of thrombosis in untreated patients is <10% per year.382383 This annual incidence translates to a risk of thromboembolism of <0.1% over the 2 or 3 days that patients are without protection. A modification of this approach, which would further decrease risk, is to delay stopping warfarin until 2 days before surgery and reverse the anticoagulant effect with a 1- or 2-mg dose of vitamin K by subcutaneous injection, repeated if the INR is still prolonged 24 hours after injection. Low doses of vitamin K1 (1 to 2 mg) have been reported to lower the INR within 24 hours without producing warfarin resistance when anticoagulant treatment is reintroduced postoperatively.384 A more aggressive approach should be considered for patients who are at high risk of developing postoperative venous thrombosis. These include patients with a past history of venous thrombosis or recurrent venous thrombosis, particularly if they have a persistent risk factor for venous thrombosis. Two treatment options are available for these high-risk patients. The first is to lower the dose of warfarin and perform the operation at an INR of ≈1.5; this approach has been shown to be safe and effective in preventing postoperative venous thrombosis in high-risk orthopedic surgical patients.385 The second option is to stop warfarin and replace the oral anticoagulant with full-dose heparin by continuous intravenous infusion preoperatively, stop heparin 6 hours before surgery, and restart anticoagulant therapy with heparin and warfarin postoperatively. Postoperative heparin should be delayed for at least 12 hours or longer if there is evidence of excessive bleeding or risk of serious postoperative bleeding. Approach to ThrombophiliaThrombophilia is defined as an increased tendency to thrombosis. It can be inherited or acquired.386 The term implies that there is an ongoing stimulus to thrombogenesis or a defect of the natural anticoagulant or fibrinolytic mechanism that predisposes the patient to thrombosis or recurrent thrombosis. Thrombophilia is often suspected on clinical grounds because a patient presents with clinical features listed in Table 10. The most important of these clinical features are idiopathic thrombosis, a family history of thrombosis, recurrent thrombosis, thrombosis at a young age, thrombosis after trivial provocation, and thrombosis in an unusual site. The spectrum of thrombophilia can range from an increased tendency to thrombosis, which is easily controlled by anticoagulant therapy, to progressive and intractable thrombosis, which is resistant to all forms of therapy. The former manifestations of thrombophilia are much more common than the latter, which is sometimes seen in malignant disease. Patients are considered thrombophilic if they have laboratory or clinical disorders that are known to be associated with an increased risk of thrombosis. Thrombophilic conditions can be inherited or acquired (Table 11). The inherited molecular abnormalities associated with an increased risk of venous thromboembolism are AT-III deficiency, protein C and protein S deficiencies, dysfibrinogenemia, and activated protein C resistance.387388389390391392393394395396 Of these, AT-III deficiency and protein C and protein S deficiency are seen in ≈10% of patients with idiopathic venous thrombosis and dysfibrinogenemia in <0.5%. Thus, until recently only ≈10% of patients with clinically suspected thrombophilia had an associated observable genetic defect. This state of affairs changed dramatically with the discovery of activated protein C (APC) resistance by Dahlback and colleagues in 1993.397 These investigators found that adding APC to plasma obtained from a patient with recurrent thrombosis failed to prolong aPTT, and the term “APC resistance” was introduced to describe the abnormality. This laboratory finding was confirmed by other investigators, who reported that between 20% and 60% of patients with recurrent thrombosis had APC resistance.398399400 APC resistance is transmitted in an autosomal dominant manner.399 The genetic defect underlying many cases of APC resistance was described in 1994. Most patients with APC resistance have a mutant factor V molecule (factor V Leiden), which resists proteolysis by APC when activated to factor Va.401 The genetic defect causing APC resistance is common; it occurs in ≈5% of normal populations.398399 Sixty percent of women in whom thrombosis occurred while they were taking oral contraceptives have been reported to have APC resistance.402 Thus, 1 of the 4 inherited disorders (APC resistance, protein C deficiency, protein S deficiency, and AT-III deficiency) is found in at least half of all patients with idiopathic venous thrombosis. The main acquired factors that predispose a patient to thrombosis are the presence of an anticardiolipin antibody (lupus anticoagulant), malignancy,182403 and chemotherapy for cancer.182 Less common are paroxysmal nocturnal hemoglobinuria, myeloproliferative disorders, nephrotic syndrome, and hormonal treatment for infertility.404405 Other inherited laboratory abnormalities for which an association has not been proved are plasminogen deficiency, heparin cofactor II deficiency, increase in histidine-rich acid glycoprotein, and reduced plasminogen activator activity due to either increased levels of plasminogen activator inhibitor or reduced levels of TPA. In addition, preoperative and postoperative reductions in fibrinolytic activity associated with increased plasminogen activator inhibitor and reduced activity of plasminogen activator have been shown to be associated with an increased risk of postoperative thrombosis.406 Investigation of ThrombophiliaLaboratory testing for inherited AT-III, protein C or protein S deficiency, and resistance to activated protein C should be performed when the patient is not being treated with anticoagulants, at a time remote from the acute thrombotic event, and after excluding the various acquired disorders known to perturb the levels of some of these naturally occurring anticoagulants. Whenever possible, the diagnosis of inherited deficiency should be confirmed by family studies. If anticoagulant therapy is indicated because of the underlying thrombotic process, then testing for AT-III deficiency can be done while the patient is being treated with oral anticoagulants (a low result would be diagnostic, although an AT-III level in the normal range would not exclude AT-III deficiency). Assays for protein C and protein S can be performed while the patient is on high-dose subcutaneous heparin. Testing for APC resistance with coagulation-based assays during anticoagulant therapy has been difficult in the past. The problem can be overcome by using genetic testing for factor V Leiden or a tissue factor–dependent factor V assay that permits reliable diagnosis of APC-resistant factor Va in patients receiving anticoagulant therapy.407 Before labeling the patient as having a deficiency, it is important to repeat the test on several occasions, exclude an underlying acquired abnormality that could produce a falsely low test result, and, if possible, perform studies in family members to confirm the inherited nature of the deficiency. If a laboratory marker of thrombophilia is found, comprehensive family studies should be performed to determine whether other family members have the defect, since asymptomatic carriers should be counseled about the need for prophylaxis when they are exposed to high-risk situations. In addition, female patients with thrombophilia or asymptomatic carriers of AT-III, protein C or protein S deficiency, and those with factor V Leiden require counseling with regard to future pregnancies, oral contraceptives, and postmenopausal estrogen replacement therapy. Many investigations for an acquired thrombophilic state can be performed at the same time as assays for inherited thrombophilia. The antiphospholipid antibody syndrome should be investigated by both an antiphospholipid antibody test and tests for circulating anticoagulants. Malignancy should be suspected in patients without other detectable causes for venous thrombosis who present with idiopathic venous thrombosis, recurrent venous thrombosis, including recurrent superficial venous thrombosis, and thrombosis in an unusual site such as the portal vein, mesenteric vein, or vena cava. Malignancy should also be suspected in patients who develop recurrent thrombosis despite adequate oral anticoagulant or heparin therapy and in patients with the syndrome of thrombophlebitis migrans. Malignancy is usually suspected on the basis of compatible clinical manifestations, although patients with occult malignancy can present with thrombosis. However, in patients with first-episode typical DVT, without special features of a thrombophilic state, expensive or uncomfortable investigations for malignant disease should not be performed if simple investigations (complete blood count, chest radiograph, and fecal occult blood testing) are negative. If malignancy is suspected, a computed tomography scan of the abdomen and/or an ultrasound of the abdomen and pelvis should be performed. If there is evidence of an iron deficiency anemia or if testing for fecal occult blood is positive, then endoscopy and/or barium studies should be performed to investigate the lower and upper gastrointestinal tract. Screening for a myeloproliferative disorder is performed by a complete blood count, including a platelet count. Paroxysmal nocturnal hemoglobinuria is suspected from the results of a complete blood count that usually show anemia with evidence of either hemolysis or bone marrow hypoplasia and the diagnosis is confirmed by demonstrating hemoglobinemia, hemoglobinuria, and hemosiderinuria and by performing Ham’s test or Hartman’s sugar/water test. The nephrotic syndrome is suspected if there is generalized edema, hypoalbuminemia, and proteinuria. Antiphospholipid SyndromeAntiphospholipid syndrome can present in a number of ways.403 Acquired circulating anticoagulants were first identified in patients with systemic lupus erythematosus in 1948,408 and an association with thrombosis was noted 15 years later.409 The phenomenon was called the lupus anticoagulant, but this anticoagulant is not restricted to patients with systemic lupus erythematosus. Antiphospholipid syndrome is diagnosed when patients with a positive assay for antibody against phospholipids have one or more of the following manifestations: venous or arterial thrombosis, thrombocytopenia, or recurrent fetal wastage.403410411412413 The laboratory tests developed to detect lupus anticoagulants include aPTT,414 kaolin clotting time, dilute phospholipid test, platelet neutralization tests, and tissue thromboplastin inhibition tests415 ; the coexistence of antibodies to cardiolipin has been noted.416 IgG and IgM antibodies to phosphatidylserine, phosphatidylinositol, phosphatidic acid, and cardiolipin were found in subjects with the lupus anticoagulant.417 Both arterial and venous thromboses occur in patients with antiphospholipid syndrome. Thrombosis in unusual sites has been described, including the Budd-Chiari syndrome, portal vein thrombosis, and inferior vena caval thrombosis. Antiphospholipid syndrome has been reported in patients with transient ischemic attacks, multi-infarct dementia, and myocardial infarction410 ; thrombosis of the cerebral, splenic, portal, hepatic, renal, subclavian, and retinal veins and of the inferior vena cava; and coumarin skin necrosis, adrenal gland hemorrhage, and infarction.210410411412413414415416417418419420421422423424425426427428429430431432 Antiphospholipid syndrome may be the cause of cerebral ischemic events in patients.433 Myeloproliferative SyndromesMyeloproliferative syndromes have been associated with cerebral or mesenteric venous thrombosis; thrombosis of the splenic vein, portal vein, and hepatic vein; and a variety of arterial events, including strokes, myocardial infarction, vague neurological symptoms, and ischemia of the fingers and toes.434435436437438439440441442 In polycythemia vera, hypercoagulability is caused by increased viscosity of blood resulting from the increased hematocrit.443 Treatment with phlebotomy should be aimed at maintaining the hematocrit in the low-normal range (40% to 45%).443 However, other agents may be needed to lower the platelet count. Anagrelide, a quinazolin compound, is a new drug available for the management of thrombocythemia. Silverstein et al444 have accumulated data on more than 500 patients with essential thrombocythemia treated with anagrelide, with a 93% success rate in reducing platelet count to less than 650×109/L. This reduction is associated with parallel relief of symptoms (transient ishemic attacks, venous thrombosis, erythromelalgia) without a significant number of side effects. Fluid retention may be a side effect of its use in some patients, including those with congestive heart failure. Anagrelide may become front-line therapy for thrombocythemic states not only because of its selective efficacy in inhibiting megakaryocyte maturation but because no documented leukemogenic effect has been found after 8 years of use. Anagrelide will not likely displace two other agents, hydroxyurea and ∝-interferon, that are effective in treatment of thrombocythemia. Even though rare cases of leukemia have followed the use of hydroxyurea in myeloproliferative disorders, the ease and effectiveness of its low-cost administration continue to make it a valuable agent for these patients. The cytokine ∝-interferon requires frequent subcutaneous injection of an expensive agent whose use may be accompanied by significant side effects. Because the side effects of interferon are usually manageable and transitory, and because injections may be decreased from three times weekly to once weekly, this cytokine should be considered for some patients. No absolute relation between platelet number and frequency of thrombotic complications has been established in thrombocythemia, but it has been demonstrated that the longer the platelet count is less than 600×109/L, the lower the incidence of thrombotic complications.445 High vascular complication rates in patients >60 years and in patients with a history of a thrombotic event make both of these patient groups candidates for therapeutic intervention. A randomized trial of hydroxyurea has demonstrated a significant reduction in thrombotic events (3.6% versus 24%) in favor of those treated with drug therapy.446 In younger patients with asymptomatic thrombocythemia, therapeutic recommendations should remain conservative.447 Low-dose aspirin appears adequate for most patients. Trousseau’s SyndromeMost patients with malignancy and thrombosis do not have Trousseau’s syndrome, which refers to a thrombus that is characteristically migratory and recurrent. Thrombosis occurs in leg veins, neck veins, superficial veins of the thorax and abdomen, the dorsal vein of the penis, subclavian and axillary veins, cerebral veins, and visceral veins.444445446 Arterial thromboses can also occur.447 Thrombosis tends to be associated with mucinogenic adenocarcinoma. Treatment of Trousseau’s syndrome can be difficult. Coumarin is usually not effective.448449 Heparin often controls the thromboembolic manifestations and can be given long term on an outpatient basis in full therapeutic doses.448 Management of ThrombophiliaAll thrombophilic patients should receive prophylaxis in high-risk situations, and some require long-term anticoagulant therapy. Lifelong anticoagulant treatment should be considered for thrombophilic patients with a documented episode of thrombosis, with or without a laboratory abnormality, while thrombophilic patients without documented evidence of thrombosis should receive prophylaxis when exposed to high-risk situations (eg, surgery, prolonged immobilization, pregnancy). In patients with polycythemia vera, the hematocrit and platelet count should be controlled with appropriate therapy. In addition, female patients with thrombophilia and asymptomatic carriers of AT-III, protein C or protein S deficiency, and the factor V gene mutation require counseling about future pregnancy, use of oral contraceptives, and postmenopausal estrogen replacement therapy. Descriptive studies in AT-III deficiency suggest that risk of thrombosis in the unprotected pregnant patient is very high.450451 While the risk of thrombosis for protein C and protein S deficiencies during pregnancy is lower than for AT-III deficiency, risk of postpartum thrombosis is high in all three groups.451 Three approaches can be used to treat a thrombophilic patient during pregnancy:
The first option should be considered for patients who must take oral anticoagulants as long as they live, including those at risk of cardiac embolism; patients with previous venous thrombosis associated with deficiencies of AT-III, protein C, and protein S or with lupus anticoagulant; and any patient with two or more previous episodes of venous thrombosis. The second option should be considered in asymptomatic carriers of AT-III deficiency and patients with previous idiopathic venous thrombosis or thrombosis during an uncomplicated pregnancy. It would be reasonable to consider either the second or third option in asymptomatic carriers of protein C or protein S deficiency and the third option in patients with only one episode of previous venous thrombosis after provocation. “Management of Deep Vein Thrombosis and Pulmonary Embolism” was approved by the American Heart Association Science Advisory and Coordinating Committee on February 15, 1996. Requests for reprints should be sent to the Office of Science and Medicine, American Heart Association, 7272 Greenville Ave, Dallas, TX 75231-4596.  Figure 1. Valve cusp thrombus (autopsy specimen). Figure 2. Cumulative incidence of recurrent venous thromboembolism after the first episode of symptomatic deep vein thrombosis. Figure 3. Location of venous thrombi in symptomatic outpatients. Reprinted from Cogo et al78 with permission. Figure 4. Diagnostic approach to deep vein thrombosis. Figure 5. Management of clinically suspected deep vein thrombosis with venous ultrasonography at presentation and on day 7. Reprinted from Prandoni et al.112 Figure 6. Diagnosis of recurrent venous thrombosis. *On venous ultrasonography; if positive in a venous segment that had been compressible on previous assessment. Figure 7. Diagnostic approach when pulmonary embolism is suspected. *Can be followed with serial venous ultrasound. +Pulmonary angiography may be preferable in a patient whose condition is unstable. xBilateral venograms could be performed initially and proceed only if results are negative. †Other combinations include low clinical probability and intermediate or indeterminant lung scans and intermediate clinical probability and low-probability lung scans.
References
|
The patient was brought into the ED with an embolism in her right pulmonary vein
zusammenhängende Posts
Urheberrechte © © 2024 frojeostern Inc.
