Deep venous thrombosis and pulmonary embolism are frequent causes of hospital admission, and an understanding of the complex mechanisms of thrombosis and thrombolysis is important in the prevention and management of patients with these disorders.

While pathologists had known for many years that thrombi at times formed in the peripheral veins, Virchow introduced the embolism concept as a result of his very cogent findings at autopsy. He observed that patients with pulmonary embolism often also had thrombi in the veins of the legs and pelvis. These thrombi were the same histological age as those found in the lungs. When thrombi were injected into the femoral veins of animals they ultimately lodged in the lungs. Virchow concluded that thrombi formed in the systemic veins and that thrombosis was caused primarily by three factors: reduced blood flow in the systemic veins, injury to the veins, and a state of hypercoagulability. These factors remain important in the pathogenesis of pulmonary embolism and are known collectively as Virchow's triad.

‘In the peripheral veins the danger proceeds chiefly from the small branches. By no means rarely do these become quite filled with masses of coagulum. As long, however, as the thrombus is confined to the branch itself, so long as the body is not exposed . . . only the greater number of the thrombi in the small branches do not content themselves with advancing up to the level of the main trunk, but pretty constantly new masses of coagulum deposit themselves from the blood upon the end of the thrombus layer after layer; the thrombus is prolonged beyond the mouth of the branch into the trunk in the direction of the current of blood, shoots out in the form of a thick cylinder farther and farther, and becomes continually larger and larger. From a lumbar vein, for example, a plug may extend into the vena cava as thick as the last phalanx of the thumb. These are the thrombi that constitute the source of real danger; it is in them that ensues the crumbling away which leads to secondary occlusion in remote vessels.’

Thrombosis is usually initiated when alterations in the vascular endothelium cause platelets to adhere to subendothelial connective tissue. Collagen binds to platelet receptors to activate enzymes that catalyse the release of arachidonic acid. Cyclo-oxygenase converts arachidonic acid to thromboxane A&sub2; which increases phospholipase C activity and stimulates platelet activation and secretion. Cyclo-oxygenase is inhibited by aspirin and non-steroidal anti-inflammatory agents. Prostacyclin, in contrast to thromboxane A&sub2;, is produced from arachidonic acid by endothelial cells and inhibits phospholipase C activity and platelet activation. Platelets release a variety of mediators after activation, including adenosine diphosphate which modifies the platelet surface, allowing fibrinogen to attach and link to adjacent platelets. Platelet-derived growth factor is also released, stimulating the growth and migration of fibroblasts and smooth muscle cells within the vessel wall.

The primary haemostatic plug is strengthened after several minutes by activation of the coagulation pathway which causes the production of thrombin and conversion of plasma fibrinogen to fibrin. The reactions require formation of a surface-bound complex and the activation of proteases by proteolysis: they are regulated by plasma and cellular cofactors and by calcium. For example, the local concentration of reactants is reduced by the flow of blood and increased by stasis. In addition, the absorption of coagulation factors to cellular surfaces and the presence of multiple inhibitors in plasma inhibit thrombosis.

Thrombolysis begins immediately after formation of the haemostatic plug. The principal physiological factor, tissue plasminogen activator, diffuses from endothelial cells to convert plasminogen absorbed to the thrombus into plasmin, which degrades the fibrin polymer. Plasmin can also degrade fibrinogen, but systemic fibrinolysis does not occur because tissue plasminogen activator activates plasminogen more effectively when it is absorbed to fibrin thrombi. Furthermore, excess plasmin is rapidly bound and neutralized by &agr;&sub2;-plasmin inhibitor, and endothelial cells also release an inhibitor of plasminogen activator.

A variety of factors may be responsible for the development of thrombosis in the absence of endothelial injury. Low blood flow increases local concentrations of coagulation reactants. This may result from venous varicosities, prolonged bed rest, or stasis during and after an operation or illness. Factors that diminish arterial blood flow include cardiac disease or shock. Congenital or acquired abnormalities in the fibrinolytic system and certain dysfibrinogenaemias also predispose to intravascular thrombosis. While the last two specific disorders are of extreme interest and help to elucidate the complex mechanisms of thrombosis formation, they are identified in only 10 per cent of patients with thrombosis.

Antithrombin forms complexes with all of the serine protease coagulation factors except factor VII. The anticoagulant effect of heparin is due to acceleration of the rate of formation of this complex. The most common antithrombin deficiency causes a mild defect in 1 of 2000 individuals. Patients with antithrombin deficiency can develop acute thrombosis or embolism which should be treated with intravenous heparin, since they usually have sufficient functional antithrombin to act as a heparin cofactor. Subsequent treatment with oral anticoagulants prevents recurrent thrombosis. Relatives of the patient should also be screened, and those found to have low antithrombin levels should receive proplylactic anticoagulation with heparin or plasma infusions prior to surgical procedures in order to raise the antithrombin III level and decrease the risk of thrombosis. Chronic oral anticoagulation is not recommended unless a clinical thrombotic episode occurs.

Protein C is converted to an active protease by thrombin after binding to thrombomodulin. Activated protein C inactivates plasma cofactors V and VIII and stimulates the release of tissue plasminogen activator from endothelial cells. The inhibitory function of protein C is enhanced by protein S. Patients with acute thrombosis and moderate protein C or S deficiency should receive heparin followed by oral anticoagulants, although treatment with warfarin may further reduce the concentration of both proteins C and S. Deficiency in proteins C or S may predispose to the rare but serious complication of warfarin-induced skin necrosis. Homozygosity for protein C deficiency is rare, but can be a cause of intravascular coagulation in the neonate. These patients may require periodic plasma infusions rather than oral anticoagulants to prevent recurrent intravascular coagulation and thrombosis. Defects in fibrinogen or plasminogen or decreased synthesis or release of tissue plasminogen activator have been described. Patients with these disorders, as well as those with abnormal plasminogen, have been successfully treated with heparin and oral anticoagulants.

In addition to the heritable disorders, many common conditions and illnesses are associated with an increased risk of thrombosis. Age is an important factor in the development of thrombosis, and pulmonary embolism primarily affects the middle-aged and elderly. Cardiac disorders, especially congestive heart failure, acute myocardial infarction, and atrial fibrillation, are particularly conducive to the development of pulmonary embolism. Metastatic malignancy, particularly carcinoma of the pancreas and prostate, is also associated with an increased incidence of pulmonary embolism. Patients undergoing major surgical procedures or suffering major trauma or burns also have an increased risk of venous thrombosis. Factors produced in damaged or ischaemic tissues or in patients with metastatic disease, together with venous stasis and endothelial injury, induce the formation of venous thrombi.

Pregnancy increases the risk of pulmonary embolism because pressure from the gravid uterus retards venous flow from the legs and pelvis. Postpartum infection may also predispose to septic thrombophlebitis and embolism, and oral contraceptives, which lower antithrombin III levels, are also associated with the occurrence of pulmonary embolism. Several haematological disorders which cause poorly defined abnormalities in circulating leucocytes and platelets predispose patients to venous thrombosis. Diseases which affect the endothelial cells, or the administration of drugs such as l-asparaginase, which inhibits production of multiple coagulation factors, may also predispose patients to thrombosis.

Haemostatic thrombi that form in veins when blood flow is reduced are richly endowed with fibrin and entrapped blood cells and contain relatively few platelets. These are often called red thrombi. The friable ends of these thrombi which form in leg veins often break off and cause emboli in the pulmonary circulation (Fig. 1) 633. Thrombus formation is typically asymptomatic and the site does not usually become inflamed. Inflammation of a thrombus in a vein is called thrombophlebitis. The acute inflammatory response makes the developing thrombus firmly adherent to the intima of the vessel wall, making embolization uncommon.

Venous thrombosis may involve the superficial or deep venous systems of the leg. When both systems are affected, thrombus formation usually begins in the deep veins and extends to the superficial system. Varicose veins are often associated with superficial thrombophlebitis of the lower extremities; other causes include occult malignant neoplasms, local trauma, and parenteral drug abuse. In a substantial number of patients the condition may be idiopathic.

The common clinical presentation of superficial thrombophlebitis is local pain, erythema, and induration, with tenderness of the thrombosed vein. When thrombophlebitis occurs below the knee, management consists of bed rest, leg elevation, and local application of heat to the affected veins. The disorder is usually self-limiting as obliteration of the affected part of the superficial venous system precludes subsequent attacks. The risk of thromboembolism is minimal and anticoagulation is not indicated. When thrombophlebitis extends above the knee, embolization may occur: such patients should be closely observed for cephalad progression of thrombus. Anticoagulation to prevent thromboembolism is indicated if the response to conservative management is poor.

Deep venous thrombosis is serious since the thrombus is much more likely to embolize to the lungs: when the thrombosis is proximal to the calf, there is a 50 per cent likelihood of pulmonary embolism, and up to 30 per cent of thrombi isolated to the calf veins embolize to the lungs. As many as 40 to 50 per cent of patients with deep venous thrombosis who develop pulmonary embolism have no symptoms of deep venous disease, causing a delay in the administration of appropriate prophylactic and therapeutic measures.

In patients who develop symptoms, mild oedema, superficial venous dilatation, and pain in the calf are usually present. Palpation of the calf may disclose tenderness and occasionally a thrombosed vein can be felt at any site from the plantar aspect of the foot to the femoral triangle in the groin. A thrombosed vein is usually best identified by palpation in the popliteal space. Homans' sign (tenderness and tightness in the calf with hyperextension of the foot) provides further evidence of thrombosis, but it may be present with any type of calf muscle irritation and is not pathognomonic for thrombotic disease.

Most forms of deep venous thrombosis involve the popliteal vein and its tributaries, but occasionally the thrombosis extends proximally to the femoral or iliac veins. Swelling and pain in the distal thigh are more prominent if femoral vein thrombosis is present, but these signs may be absent. Phlegmasia caerulea dolens is the condition found when ileofemoral thrombosis is associated with massive swelling of the entire extremity to the inguinal ligament, severe pain, tenderness, and cyanosis. Ileofemoral arterial thrombosis with spasm is frequently present and is characterized by a pale cool extremity with diminished or absent pulses. Disease confined to the popliteal vein and its tributaries may be occult or confused with other conditions such as rupture of the gastrocnemius muscle or disorders involving the knee, particularly a ruptured Baker's cyst. It is therefore important to confirm objectively the presence of suspected deep venous thrombosis.

The most specific test for confirmation of the diagnosis of deep venous thrombosis is venography, in which contrast medium is injected into a vein on the dorsum of the foot to demonstrate the venous drainage through the popliteal, femoral, and iliac veins (Fig. 2) 634. A normal result nearly always excludes the presence of venous thrombosis. However, venography may be complicated by venous thrombosis and extravasation of contrast media produces perivasculitis, cellulitis, and occasionally ulceration of the skin. Patients are, therefore, often initially screened with a non-invasive technique and undergo venography if the diagnosis remains in doubt. Venography using radioisotopes instead of contrast medium avoids complications, but although results with this technique are improving it is not in widespread use.

Real-time B-mode ultrasonic imaging combined with Doppler ultrasound (duplex scanning) is a practical, non-invasive method of assessment of blood flow in veins and valve cusp movement, and can differentiate between acute and chronic thrombosis. All of the major deep veins of the lower limb can be assessed; however, it cannot exclude the presence of thrombi in small veins and is less accurate in demonstrating thrombotic disease in the calf. With experience, it is both accurate and relatively inexpensive and can be used in patients requiring reassessment. Plethysmography is another non-invasive technique which is useful in the diagnosis of deep venous thrombosis. A calf plethysmograph measures volume changes and may detect the oedematous changes associated with thrombus formation. Like duplex scanning, plethysmography is helpful in demonstrating proximal thrombotic disease than that occurring in the calf.

Intravenous administration of radioactive fibrinogen is another sensitive non-invasive technique used to diagnose deep venous thrombosis. Following intravenous injection of [¹²&sup5;I]fibrinogen, the legs are scanned with a gamma-camera to detect the developing thrombus, which incorporates radioactive fibrinogen. This test is particularly accurate for detecting thrombosis in the calf, but high background radiation from the pelvic bones and urinary bladder means that it is not useful in assessing veins in the upper thigh. Any inflammatory condition also causes increased radioactivity, and results are not available for at least 12 h. Radioactive scanning of the lower extremities demonstrates deep venous thrombosis in as many as 54 per cent of patients following surgical treatment of a fractured hip, 50 per cent of patients following prostatectomy, and 28 per cent of general surgical patients over the age of 40. Scanning following injection of ¹¹¹I-labelled platelets may also be useful in the diagnosis of deep venous thrombosis.

Magnetic resonance imaging (MRI) is a reliable method of diagnosing venous thrombosis and can demonstrate thrombi in the pelvic veins (Fig. 3) 635. MRI does not require the administration of intravenous contrast agents and can be used safely in patients with allergies to dye or with impaired renal function.

Anticoagulation prevents the propagation of the original thrombus and the development of new thrombi while the existing thrombus is lysed by naturally occurring fibrinolysis. Intravenous heparin has rapid action and can be discontinued, or its effects reversed rapidly with protamine sulphate, to decrease the possibility of bleeding complications. Subcutaneous heparin is also effective.

Although anticoagulation is management of choice for most patients, fibrinolytics such as urokinase, streptokinase, and recombinant tissue plasminogen activator have been evaluated clinically. Fibrinolytics can completely lyse up to 70 per cent of existing thrombi, a feature that conventional anticoagulants lack. Among 108 patients with venographically verified deep venous thrombosis treated with streptokinase, total or partial thrombolysis was demonstrated angiographically in 60 (55.6 per cent). However, three died during treatment, all from pulmonary embolism, and six developed clinical signs suggestive of pulmonary embolism. Major bleeding was a complication in 16 patients (14.8 per cent), allergic reactions to streptokinase occurred in 23 patients, and one patient developed anaphylactic shock. Thus, although streptokinase was effective in the management of deep venous thrombosis, complications were significant and the routine use of fibrinolytic therapy has been prohibited by bleeding complications, especially in postoperative situations. In addition, lysis of thrombi more than 72 h old is reduced and fibrinolytics have no advantage over heparin in the prevention of recurrent venous thrombosis. Their role in managing deep venous thrombosis is therefore limited, although patients with extensive disease, such as in the iliofemoral system, may benefit from their use.

The surgical extraction of venous thrombi has been almost completely discontinued since the recurrence rate is high. Venous thrombectomy still has a role in the management of patients with extensive iliofemoral disease in which limb loss is imminent, such as in phlegmasia alba dolens.

With the increasing use of the intravenous route for infusion of fluid and pharmacological agents, the incidence of thrombophlebitis in the upper extremity has increased. Superficial thrombophlebitis is often associated with an indwelling venous catheter, and treatment is similar to that of superficial phlebitis in the lower extremity. If infection is suspected, the vein is aspirated with a large bore needle: if suppurative material is obtained the vein should be surgically excised. Antibiotics should be given, but anticoagulant therapy is usually not necessary since pulmonary embolism is an uncommon complication.

Thrombosis of the axillary and/or subclavian vein may occur when the upper extremity has been vigorously abducted at the shoulder, ceasing trauma to the vein. Venous obstruction may also occur when the vein is compressed between the clavicle and first rib, producing non-pitting oedema of the entire arm. Cyanotic mottling of the skin and distension of the superficial veins are also present. Axillary vein thrombosis may also be caused by congestive heart failure, indwelling venous catheters, external trauma, and neoplastic disease in the region of the axilla. The diagnosis is usually obvious after the history and physical examination and can be confirmed by venography. Treatment consists of arm rest in an elevated position and administration of anticoagulants. Venous thrombectomy may occasionally be performed, but rethrombosis is common. Subclavian vein occlusion may occur as an extension of axillary vein thrombosis or as a separate entity, and is commonly due to an indwelling venous catheter. Management is similar to that of axillary vein thrombosis. Recent series suggest that pulmonary embolism occurs in up to 12.4 per cent of patients with deep venous thrombosis of the upper extremity.

Occlusion of the superior vena cava may be caused by compression from a neoplasm or by lymphadenopathy. Trauma and/or infection from a central venous catheter may also cause thrombosis at this site. The superior vena caval syndrome includes cyanosis of the head and neck, with oedema of the upper chest and extremities and varying degrees of venous distension. Patients may complain of vertigo, headache, epistaxis, and occasionally fainting. Venography or MRI confirms and defines the level and extent of venous obstruction. Treatment includes anticoagulation and specific treatment of the underlying condition. Surgical excision of a neoplastic mass is occasionally indicated, and radiation therapy may be successful.

Pulmonary embolism is a common and sometimes fatal complication of deep venous thrombosis. Although it is recognized in the postoperative period, most patients develop pulmonary embolism secondary to non-surgical disorders, including congestive heart failure, cerebrovascular accidents, chronic pulmonary disease, systemic infections, carcinomatosis, and many chronic disorders.

Emboli that prove fatal are generally 1.5 cm or more in diameter and 50 cm or more in length, and are often fragmented (Fig. 4) 636. The right pulmonary artery is more commonly affected than the left, and the lower lobes more often than the upper lobes. Emboli originate primarily in the systemic venous circulation—most arise in the iliac and femoral veins, but up to 20 per cent originate from other sources, including the inferior vena cava, the subclavian, axillary, and internal jugular veins, and occasionally the cavernous sinuses of the brain. Emboli due to neoplasms should also be considered in the differential diagnosis. Renal cell carcinoma metastasizes early in its clinical course, and direct extension to the renal vein and inferior vena cava may cause pulmonary embolism in 10 to 54 per cent of affected patients. Primary pulmonary neoplasms can also mimic pulmonary embolism. Cardiac tumours arising in the right atrium and right ventricle may be the site of extensive pulmonary emboli.

A primary feature of pulmonary blood flow is the low vascular resistance that enables flow in this bed to increase several-fold with minimal elevation of pulmonary arterial pressure. Physiological changes following pulmonary embolism are related to the size of the emboli and can be divided into those that produce microembolism (obstruction of terminal small arteries and arterioles) and those that produce macroembolism (occlusion of the large pulmonary vessels). Considerable reduction in the diameter of the main pulmonary artery and the primary branches (at least 50 per cent) is required to reduce pulmonary blood flow significantly or to produce pulmonary hypertension. Experimental thrombi of large diameter produced in the inferior vena cava and embolized to either the right or the left pulmonary artery 10 to 14 days after their formation produce minimal cardiovascular and respiratory responses. Occlusion of one pulmonary artery usually causes insignificant changes in the central venous pressure, right ventricular pressure, pulmonary arterial pressure, systemic arterial pressure, cardiac output, total oxygen consumption, and the electrocardiogram, despite occlusion of half of the pulmonary arterial circulation, provided the remaining lung is normal. If one lung is normal or nearly normal, removal of the opposite lung is relatively well tolerated: tidal volume and oxygen consumption at rest change only minimally. Similarly, ligation of one pulmonary artery or occlusion by an intraluminal balloon is accompanied by few cardiodynamic changes. During exercise similar occlusion causes an increase in pulmonary arterial pressure of only 12 to 50 per cent, while cardiac output may increase as much as three-fold. Such occlusion closely simulates the obstruction produced by large pulmonary emboli. It should be emphasized that these findings apply to healthy subjects: underlying cardiac or respiratory insufficiency alters this response appreciably. In patients with heart disease, exercise during unilateral occlusion of the right and left pulmonary artery by a balloon catheter produces a sharp elevation in pulmonary arterial pressure. Resection of less than one lung is followed by only minor changes in the pulmonary arterial pressure, whereas removal of a greater amount of pulmonary tissue is associated with elevated pulmonary arterial pressure.

Clearly, mechanical factors are the most important in determining the cardiodynamic effects of pulmonary embolism, but reflex effects may cause bronchoconstriction. Tachypnoea, pulmonary hypertension, and systemic hypotension have also been demonstrated following experimental embolization with small particles. This is probably not a common clinical problem, although it may occur after massive blood transfusion during which small emboli containing platelets, leucocytes, and fibrin may occlude the pulmonary microcirculation.

A clinical diagnosis of pulmonary embolism may be difficult because of its similarity to a number of other cardiorespiratory disorders. Dyspnoea, chest pain, and haemoptysis are classic symptoms but are not sufficiently specific to establish a definite diagnosis. It should be emphasized that many patients have underlying cardiac disease, and dyspnoea and tachypnoea are the most frequent clinical findings. Accentuation of the pulmonary second sound is also common. Haemoptysis, pleural friction rub, gallop rhythm, hypotension, cyanosis, and chest splinting are present in no more than one-quarter of patients. Clinical evidence of venous thrombosis occurs in only one-third of patients. The signs and symptoms found in 1000 consecutive patients are shown in Table 1 215.

In patients with acute pulmonary embolism and no other pulmonary disease, the plain chest radiograph is usually normal. Diminished pulmonary vascular marking at the site of embolism may be present (Westermark's sign). The ECG is not specific but significant changes can be confirmatory. No more than 10 to 20 per cent of patients subsequently proven to have pulmonary embolism show any alterations in the ECG, including disturbance of rhythm (atrial fibrillation, ectopic beats, heart block), enlargement of P waves, S-T segment depression, and T-wave inversion (particularly in leads III, AVF, V&sub1; and V&sub3;), and of these, a smaller number show diagnostic abnormalities. The most common abnormality is S-T segment depression due to myocardial ischaemia, reduced cardiac output, and low systemic arterial pressures, as well as increased right ventricular pressure. Arterial blood gases are often normal, especially in young patients. Because clinical findings and routine examinations are non-specific, radioactive pulmonary scanning and pulmonary angiography are used to diagnose pulmonary embolism (Fig. 5) 637. A schematic outline of the plan to be followed in establishing the diagnosis is shown in Fig. 6 638.

Radioisotope perfusion and ventilation pulmonary scans remain the most frequently employed technique in the diagnosis of pulmonary embolism. The method involves the detection of intravenously injected particles such as technetium–99m that become lodged in the pulmonary capillary bed.

The results of ventilation/perfusion scans are usually described in terms of the probability of embolism. High probability is indicated by segmental or greater perfusion defects with a normal ventilation scan (V/Q mismatch). Moderate probability is indicated by multiple subsegmental perfusion defects with a normal ventilation scan or segmental perfusion defects without a ventilation scan. Scans which are indeterminate show chronic obstructive pulmonary disease on the chest film or regions of perfusion scans, while subsegmental perfusion defects without a ventilation scan or matched perfusion and ventilation defects indicates a low probability of an embolism. High probability scans have a high positive predictive value for pulmonary embolism while low probability scans with low clinical suspicion are rarely associated with such emboli. Intermediate probability scans and low probability scans with high clinical suspicion are often associated with pulmonary embolism and indicate the need for arteriography to be performed.

Venograms or other non-invasive tests may be useful in establishing a definitive diagnosis of deep venous thrombosis, particularly in patients in whom the diagnosis is in doubt or when insertion of a vena caval umbrella is being considered.

The most definitive test for the diagnosis of pulmonary embolism is pulmonary arteriography. It is important that the appearance of the normal pulmonary angiogram is familiar in order that morphological and physiological changes can be appreciated. The arteries in the lower areas of the lung are normally larger because of a greater volume of pulmonary tissue. In most patients who survive the initial embolic episode the obstruction in the pulmonary arteries involves lobar or segmental branches. The defect should remain constant on several successive films in the series, and the flow may be sluggish, shown by a small pool of contrast medium that may persist in the artery above the obstruction after the venous phase of the angiogram. When pulmonary arteriography is performed later in the course of embolism, contrast medium may pass around the obstruction, causing delayed opacification of the artery distally. In some areas avascular segments resulting from unresolved thromboembolism may be seen. Oblique views of the pulmonary arteriogram should be obtained for maximal visualization and diagnostic accuracy. Pulmonary arteriography is safe when experienced radiologists use small diameter catheters and low osmolarity contrast agents.

The importance of reaching an unequivocal diagnosis of pulmonary embolism before therapy is instituted warrants emphasis. The most reliable means of achieving absolute diagnosis are pulmonary scanning and arteriography. Because scanning is a simple, safe, and reliable technique, it is generally performed initially. If the pulmonary scan is to be used for definitive diagnosis, a concomitant plain chest film must show a normal pulmonary appearance in the area in which the scan demonstrates pulmonary arterial occlusion.

Prophylaxis of deep venous thrombosis and pulmonary embolism is an important aspect of postoperative care. Nevertheless, no method or combination of methods completely prevents thromboembolism. Factors which reduce the risk include physical activity and elevation of the lower extremities for gravity drainage of venous return. Some consider compression of the legs by stockings or mechanical devices and prophylactic anticoagulation to be useful, but many disagree and these issues remain controversial. Early ambulation and resumption of physical activity after operation or bed rest for any reason has long been recommended.

Antiplatelet drugs play a role in the management of patients with thromboembolism. Non-steroidal anti-inflammatory drugs, including aspirin and dipyridamole, have also been shown to inhibit the platelet-release reaction secondary to ADP-induced platelet aggregation and adherence to collagen in vitro. Dipyridamole inhibits phosphodiesterase and raises intracellular cyclic AMP levels. The usual dose of 50 to 100 mg four times daily has no effect on platelet function, and it is usually administered in a combination with aspirin. Aspirin reduces the incidence of thrombosis from 20.4 per cent in controls to 12.5 per cent. In one study 1.2 g of aspirin daily was shown to be as effective as warfarin in preventing clinically diagnosed venous thrombosis and pulmonary embolism in a group of patients undergoing elective hip replacement. The incidence in the group receiving aspirin was 9 per cent, compared with 35 per cent in a previously studied control group. Studies of the overall effectiveness of antiplatelet drugs in the prevention of venous thrombosis and pulmonary embolism are continuing, because absolute certainty concerning their use has not yet been established. Dextran has also been used as an antiplatelet drug to prevent deep venous thrombosis.

Elevation of the legs with flexion of the knee as depicted in Fig. 7 639 causes a rapid runoff of the blood in the veins of the leg and thigh due to gravity. This is a simple, effective, and broadly applicable prophylactic measure. Pneumatic compression devices also decrease stasis and increase venous blood flow but are not widely used.

While prophylactic anticoagulation may be beneficial, especially after trauma and in patients with orthopaedic disorders, the concept of low-dose heparin as a prophylactic measure remains controversial. The usual regimen is an initial dose of 5000 units subcutaneously, repeated every 8 to 12 h until the patient is fully ambulatory. Coagulation times are prolonged minimally, if at all, with a low risk of bleeding. The protection conferred, if any, may be due to the potentiation of a naturally occurring plasma inhibitor of activated factor X. A large number of trials with low-dose heparin given to surgical patients postoperatively have used ¹²&sup5;I-labelled fibrinogen scanning or venography, or both, for the demonstration of development of venous thrombosis. With some exceptions, these studies have generally indicated a decrease in the occurrence of deep venous thrombosis, compared with that in controls. The efficacy of low-dose heparin in the prevention of postoperative pulmonary embolism is less obvious. The most frequently quoted of these studies is a multicentre clinical trial involving more than 4000 patients. Another randomized study of a series of patients aged over 40 years undergoing intraperitoneal procedures under general anaesthesia lasting more than 30 min found that the incidence of calf vein thrombosis was reduced by low-dose heparin but that the incidence of proximal vein thrombosis and pulmonary emboli, which were detected by chest films, pulmonary function tests, and perfusion scanning, was not reduced.

Although low-dose heparin continues to be recommended by some to prevent thromboembolism, it is not frequently used. It appears to have limited value after prostatectomy, after myocardial infarction, and in major orthopaedic procedures, particularly repair of femoral fractures and reconstructive surgery of the hip and knee. Low-dose heparin prophylaxis is also inadequate for patients with an active thrombotic process. Finally, it may lead to heparin sensitivity and cause disseminated intravascular coagulation, a condition in which platelets aggregate into thrombi and which may result in gangrene of the extremities.

Anticoagulation with heparin is the standard treatment for acute thrombosis and pulmonary embolism. Heparin is administered intravenously at an initial dose of 5000 to 10 000 units followed by constant infusion at a rate sufficient to raise the activated partial thromboplastin time to 1.5 to 2 times the control value, usually 1000 units/h. Since heparin is excreted mainly in the urine, the patient's renal status must be monitored. The duration of heparin therapy varies, but 5 to 10 days is generally appropriate as this is the time usually required for thrombi to become adherent to the venous wall. Oral coumarin therapy, begun several days before cessation of heparin therapy, allows adequate prolongation of the prothrombin time.

The major complication associated with heparin therapy is bleeding, especially from surgical sites and into the retroperitoneum. Aspirin therapy, poor platelet function, and intramuscular injections may contribute to the risk of significant bleeding. Although heparin anticoagulation can be rapidly reversed by the administration of protamine sulphate, this is not usually necessary as reduction or cessation of heparin often stops the bleeding. Delayed haemorrhage may occur in patients with recent prosthetic arterial grafts. Continuous lysis and resorption of old thrombus and its replacement with new thrombus occurs in suture lines until they are sealed by regeneration of new intima. Haemorrhage may occur up to 1 month after the placement of arterial grafts in patients maintained on heparin therapy. Contraindications to heparin therapy include internal bleeding, intracranial neoplasm, recent cranial surgery, trauma, or haemorrhagic stroke.

Thrombocytopenia occurs in about 10 per cent of heparin recipients: this may be severe, accompanied by intravascular platelet aggregation and arterial thrombosis. Recognition of this complication is critical since cessation of heparin can reverse this syndrome and may be life-saving. Commercial heparin preparations are heterogeneous and only about 20 per cent of the infused material has anticoagulant activity. Low molecular weight heparin fractions that retain anticoagulant activity are the treatment of choice in patients with heparin-dependent thrombocytopenia who require additional therapy. These fractions do not interact with platelets and may not cause thrombocytopenia. Administration of heparin for longer than 2 months is also associated with a risk of osteoporosis.

The coumarin anticoagulants (coumadin and warfarin), which prevent the reduction of vitamin K in the liver and induce a state analogous to vitamin K deficiency, are used to prevent the recurrence of venous thrombosis and pulmonary embolism. The average loading dose of coumadin is 15 to 30 mg on the first day and 10 to 20 mg on the second day. The maximal effect is usually reached in 36 to 48 h, and the average daily maintenance dose is usually between 5 and 10 mg (range 2–20 mg). Anticoagulation is easily achieved by administration of coumadin once daily, adjusting the dose until the desired prolongation of the prothrombin time is achieved.

The prothrombin time should be monitored regularly in patients receiving oral anticoagulants. Despite the most careful management, prothrombin time may fluctuate, and drugs that alter anticoagulant metabolism or degradation in the liver or which compete with albumin binding sites can increase or decrease their potency.

There is a direct relationship between the duration of anticoagulation and the risk of recurrent thrombosis. Although recommendations vary, most patients with a single uncomplicated thromboembolic event have maximal benefit after 3 to 6 months of anticoagulation. About 10 per cent of patients treated with an oral anticoagulant for 1 year have a serous complication requiring medical supervision and 0.5 to 1 per cent have a fatal haemorrhagic event, despite careful medical management. The anticoagulation effects of coumadin can be reversed by infusion of fresh frozen plasma or by the administration of vitamin K. In many patients reduction or omission of several doses improves haemostasis and stops haemorrhage. Despite the risk of bleeding, patients with prosthetic heart valves, severe mitral stenosis, cardiomyopathy, chronic congestive heart failure, recurrent or persistent atrial fibrillation, and an inherited prethrombotic disorder may require life-long anticoagulation.

Activators of the fibrinolytic system are frequently used to accelerate lysis of thrombi. These agents are either naturally occurring products or chemically modified derivatives, and they differ with respect to fibrin specificity and complications. Current indications for fibrinolytic therapy include patients with massive pulmonary embolism complicated by hypotension, severe hypoxaemia, and right heart strain or failure. In addition, fibrinolytic agents have been successfully administered to patients with extensive iliofemoral thrombophlebitis. While such therapy often resolves venous thrombi, there is no firm evidence that lytic therapy reduces the incidence of long-term complications. Fibrinolytic therapy may be of benefit in patients with thrombosis of the axillary vein, which does not respond well to conventional anticoagulation.

Two thrombolytic agents, streptokinase and urokinase, have been studied extensively. Both act by transforming plasminogen to plasmin but cannot discriminate between free and fibrin-bound plasminogen. Streptokinase is a soluble product of the metabolism of Streptococcus pyogenes (Lancefield Group A) which indirectly activates plasminogen. Patients who have suffered previous streptococcal infections may be allergic to streptokinase, clinically manifest as toxic reactions such as pyrexia, dyspnoea, tachycardia, and anaphylaxis. Urokinase is a product of renal epithelial and tubular cells which directly activates plasminogen. In a national co-operative study, urokinase combined with heparin therapy, compared with heparin alone, significantly accelerated the resolution of pulmonary thromboemboli at 24 h, as shown by pulmonary arteriograms, pulmonary scans, and right-sided heart pressure measurement. However, no significant differences in the recurrence rate of pulmonary embolism or in 2-week mortality were noted. Bleeding was a common complication, occurring in 45 per cent of patients receiving urokinase and heparin, compared with 27 per cent of those given heparin alone.

In a study of the long-term effects of thrombolytic treatment of acute and massive embolism, seven patients underwent pulmonary angiography with pressure measurements before and after intrapulmonary infusion of urokinase (average dose, 1724 units/kg.h) and heparin (average dose 17 units/kg.h). The treatment was monitored by daily measurements of pulmonary arterial pressure and was continued until normal pressure was achieved (an average of 6 days). Pulmonary angiograms showed massive obstruction before therapy, with improvement occurring within 6 days after treatment. The mean pulmonary arterial pressure declined from an average of 37 &plusmin; 9 to 13 mmHg after 6 days and to 15 &plusmin; 3 mmHg after 15 months. No recurrence of pulmonary embolism was observed. Mean pulmonary arterial pressure and total pulmonary resistance remained within normal limits in six of seven patients, at rest and during bicycle exercise in the supine position. All patients showed clinical signs of deep venous thrombosis early after treatment. Fifteen months later, four patients had normal deep veins, and three had phlebographic signs of old thrombosis. Normal pulmonary arteriograms were obtained in six of seven patients. The reserve capacity of the pulmonary vasculature assessed during heavy exercise was normal.

Streptokinase is usually given with a loading dose of 250000: this may need to be repeated since patients may have antistreptococcal antibodies. Urokinase is given with a loading dose of 4400 units/kg body weight, administered over 10 to 30 min. A systemic lytic state develops with a decrease in fibrinogen levels, prolongation of the thrombin time, and prolongation of the euglobin lysis time. After the initial loading dose, 100000 units of streptokinase or 4400 units of urokinase/kg body weight are administered hourly for 24 to 72 h. The lytic state is reversed by discontinuing the enzyme. Heparin therapy can be started after 6 h and is continued for 7 to 10 days. Fibrinolytic therapy should be initiated as soon as possible after the onset of thrombosis or embolism. The systemic fibrinolysis associated with the fibrinolytic agents may cause haemorrhage since essential haemostatic plugs are attached as well as pathological thrombi. Lytic therapy is therefore not recommended for patients with recent surgery or those with a history of a neurological lesion, gastrointestinal bleeding, or hypertension.

Recombinant tissue plasminogen activator is now available for general use. In one study, a group of patients with angiographically documented pulmonary embolism, all of whom had segmental or proximal pulmonary arterial obstruction within 5 days of the onset of symptoms or signs, were treated with 50 mg of recombinant tissue plasminogen activator every 2 h followed by an additional 40 mg every 4 h if required. Thirty-four of the 36 patients had angiographic evidence of thrombolysis by 6 h: clot lysis was slight in four, moderate in six, and marked in 24. Fibrinogen levels decreased by 30 per cent at 2 h and by 38 per cent at 6 h, with only two major complications. Infusion of recombinant tissue plasminogen activator via the artery does not offer significant benefit over administration by the intravenous route. Firm contraindications to the use of thrombolytic therapy include internal bleeding (recent or active), recent neurosurgery, insertion of an arterial prosthetic graft, cranial trauma, and a history of haemorrhagic cerebrovascular accident. Relative contraindications include a recent surgical procedure (within 7–10 days), cardiopulmonary resuscitation (within 7–10 days), or the presence of a coagulopathy.

While anticoagulant therapy of pulmonary embolism is usually successful, it may fail, and in this event the need for surgical management should be reviewed on an individual basis. Venous thrombectomy was previously recommended but is now rarely performed because of the high incidence of recurrent postoperative thrombosis. The presence of phlegmasia caerulea dolens with secondary arterial spasm is a rare indication for thrombectomy. Although thrombosis may recur in such patients, the venous lumen may remain patent for long enough to relieve the arterial spasm and prevent gangrene developing.

Although vena caval interruption was previously recommended for selected patients with pulmonary embolism, it is seldom performed today. A stainless steel umbrella designed by Greenfield and Michna can be inserted under local anaesthesia through the femoral or jugular vein. With this device a filter is fixed to the wall of the inferior vena cava by hooks (Fig. 8) 640. Complications include distal migration to the bifurcation of the inferior vena cava, protrusion of the struts through the caval wall, formation of thrombus on the filter, misplacement of the device, retroperitoneal haemorrhage, perforation of the duodenum or ureter, and development of a thrombus proximal to the umbrella, producing emboli. The filter may also migrate into the iliac vein, renal vein, right atrium, right ventricle, or pulmonary artery, and such migration is occasionally fatal. The filter may also stimulate distal thrombosis in the vena cava and late occlusion may occur. Other filter types used are the Amplatz, Gunther, and birds nest.

In 1908, Trendelenburg performed the first pulmonary embolectomy. He treated three patients with this procedure, the longest survival time being 36 h. In 1924, Kirschner performed the first successful pulmonary embolectomy associated with long-term survival. The first successful pulmonary embolectomy performed using extracorporeal circulation was reported in 1961: this is currently the preferred technique, since it permits continuous oxygenation of the body while the emboli are safely removed from the pulmonary arteries.

Persistent and refractory hypotension despite maximal resuscitation is the primary indication for pulmonary embolectomy, especially in a patient with massive embolism clearly documented by either a pulmonary scan or pulmonary arteriogram. Treatment includes systemic heparinization and the administration of vasopressors, inotropic agents, and endotracheal oxygen. The primary management should be by this approach, since many patients previously thought to require embolectomy respond favourably with intensive resuscitation. Usually, 1 or 2 h may be spent attempting to restore acceptable cardiopulmonary function, unless the clinical situation is desperate. An appropriate blood pressure should be maintained and embolectomy may be postponed, especially if renal and cerebral function is acceptable.

If pulmonary embolectomy is indicated, it is usually performed using cardiopulmonary bypass. A median sternotomy is made for exposure of the pulmonary artery. The main pulmonary artery is usually found to be free of emboli, although partial obstruction may be present. The emboli are removed from the right and left pulmonary arteries and from their major branches. Smaller emboli may be removed by passage of a Fogarty catheter and irrigation with saline. The pulmonary artery is closed and cardiopulmonary bypass is gradually discontinued, allowing the heart and lungs to resume normal function. An illustration of an embolectomy specimen and the patient's scan is shown in Fig. 9 641.

Patients with acute and severe cardiopulmonary collapse can be supported by partial cardiopulmonary bypass by a circuit from femoral vein to femoral artery for immediate resuscitation. If extracorporeal circulation is not available, a right or left thoracotomy with exposure of the most severely involved pulmonary artery can be performed, the side of predominant occlusion being determined by a scan or arteriogram. An anterior thoracotomy is appropriate for exposure of either the right or left pulmonary artery, which can be opened distally for removal of emboli while normal circulation to the opposite lung is maintained. A serious complication which may follow pulmonary embolectomy is massive endobronchial haemorrhage. Successful management involves endotracheal intubation for selective collapse of the lung and entrapment of the haemorrhage into the involved lung. Reperfusion pulmonary oedema may occur after pulmonary artery thromboendarterectomy, and prolonged mechanical ventilation is often required. The syndrome is a cause of postoperative hypoxaemia with local pulmonary infiltrate.

Most pulmonary emboli eventually resolve as a result of the action of the natural fibrinolytic systems. In a small number of patients, however, these emboli gradually accumulate in the pulmonary arterial system owing to inadequate fibrinolysis or recurrent episodes of embolism. Patients with chronic pulmonary emboli have a history of exertional dyspnoea progressing to severe respiratory insufficiency over several months or years. They may also complain of recurrent episodes of thrombophlebitis, haemoptysis due to the presence of large bronchial collaterals, and chest pain. Physical findings include signs of severe pulmonary hypertension, often combined with evidence of right ventricular failure; this may be manifested as a increased pulmonary second sound, a systolic murmur, hepatomegaly, and a S3 or S4 gallop. Medical management is usually unsatisfactory, and these patients have a poor prognosis.

Chest radiographs show a dilated pulmonary artery and oligaemic pulmonary fields in approximately half of these patients. Right ventricular enlargement is present in two-thirds and pleural effusion in approximately one-third of patients. Analysis of arterial blood gases in patients breathing room air reveals evidence of severe respiratory insufficiency, with hypoxaemia and arterial oxygen tension values of 55 to 60 mmHg and an arterial carbon dioxide tension (Paco&sub2;) of approximately 30 mmHg. The electrocardiogram is usually suggestive of chronic cor pulmonale, including right-axis deviation and right ventricular hypertrophy. Peripheral venography or MRI demonstrates venous thrombosis and indicates the source of the emboli.

Ventilation and perfusion radionuclide scans are consistent with the presence of pulmonary emboli, and perfusion defects correspond to oligaemic regions of the plain chest film and pulmonary arteriogram. Perfusion defects are usually noted bilaterally. Pulmonary arteriography allows documentation of emboli, determination of anatomical distribution of emboli, and recording of pulmonary artery pressure: the natural history is related to the magnitude of pulmonary arterial hypertension. If the mean pulmonary artery pressure is more than 30 mmHg, survival at 5 years is only 30 per cent, while only 10 per cent of those with mean pressure greater than 50 mmHg are alive at 5 years. Arteriography usually shows emboli in both lungs, with 55 to 75 per cent of the total pulmonary blood flow obstructed. Further preoperative studies include a thoracic aortogram with selective bronchial arteriography to demonstrate dilated and tortuous bronchial vessels. The bronchial circulation is often considerably dilated and communicates by collaterals with the distal pulmonary arteries.

Embolectomy may be performed on one or both pulmonary arteries. Either a right or left anterior thoracotomy can be undertaken when there is proximal occlusion of one vessel. Patients with bilateral pulmonary emboli, or with embolus of the main pulmonary artery, generally requires extracorporeal circulation during the procedure. These emboli are firmly attached to the artery wall and great care is required in the dissection. All distal emboli should be removed until there is adequate back-bleeding of bright red arterialized blood. Satisfactory distal back-bleeding can usually be predicted in advance from the information gained by selective injection of the bronchial arteries.

Postoperative complications include right ventricular failure in patients with long-standing cor pulmonale and pulmonary hypertension, haemorrhagic lung syndrome, which can be managed successfully by tracheal intubation with a dual lumen catheter and balloon occlusion of the affected lung, and phrenic nerve paresis, usually as a result of topical hypothermia. Psychiatric disturbances may also occur and are usually transient.

Embolectomy for chronic pulmonary embolism generally decrease pulmonary artery pressures and increases Pao&sub2; toward normal. In patients with proximal pulmonary arterial obstruction pulmonary embolectomy is likely to produce relief of respiratory insufficiency, a reduction in pulmonary hypertension, and an improvement of right-sided heart failure.

Bauer KA, Rosenberg RD. Congenital antithrombin III deficiency: insights into the pathogenesis of the hypercoagulable state and its management using markers of hemostatic system activation. Am J Med 1989; 87 (Suppl 3B): 39S–43S.

Bettman MA. Noninvasive and venographic diagnosis of deep vein thrombosis. Cardiovasc Intervent Radiol 1988; 11: S15.

Broekmans AW, Vektkamp JJ, Bertina RM. Congenital protein C deficiency and venous thromboembolism: a study of three Dutch families. N Engl J Med 1983; 309: 340.

Chitwood WR Jr, Lyerly HK, Sabiston DC Jr. Surgical management of chronic pulmonary embolism. Ann Surg 1985; 201: 11.

Coon WW. Venous thromboembolism: prevalence, risk factors, and prevention. Clin Chest Med 1984; 5: 391.

Ezekowitz MD, et al. Indium–111 platelet scintigraphy for the diagnosis of acute venous thrombosis. Circulation 1986; 73: 668.

Hirsch J. Diagnosis of venous thrombosis and pulmonary embolism. Am J Cardiol 1990; 65: 45C.

Huisman MV, et al. Unexpected high prevalence of silent pulmonary embolism in patients with deep vein thrombosis. Chest 1989; 95: 498.

Kelley MA, Carson JL, Palevsky HI, Schwartz JS. Diagnosing pulmonary embolism: new facts and strategies. Ann Intern Med 1991; 114: 300.

Latham B, Kafoy EA, Barrett OJr, Gonzalez MF. Deficient tissue plasminogen activator release and normal tissue plasminogen activator inhibitor in a patient with recurrent deep vein thrombosis. Am J Med 1990; 88: 199.

Loscalzo J, Braunwald E. Tissue plasminogen activator. N Engl J Med 1988; 319: 925.

Makhoul RG, Greenberg CS, McCann RL. Heparin-associated thrombocytopenia and thrombosis: A serious clinical problem and potential solution. J Vasc Surg 1986; 5: 522.

Maxwell RJ, Greenfield LJ. Effects of pulmonary embolism on survival of patients with Greenfield vena caval filters. Surgery 1987; 101: 389.

Sabiston DC Jr, Wagner HN Jr. The diagnosis of pulmonary embolism by radioisotope scanning. Ann Surg 1964; 160: 585.

Sabiston DC Jr, Wolfe WG. Experimental and clinical observations on the natural history of pulmonary embolism. Ann Surg 1968; 168: 1.

Sandler DA, et al. Diagnosis of deep vein thrombosis: comparison of clinical evaluation, ultrasound, plethysmography, and venoscan with X-ray venogram. Lancet 1984; ii: 716.

Schafer AL. The hypercoagulable states. Ann Intern Med 1985; 102: 814.

Thomas DP. Overview of venous thrombogenesis. Sem Thromb Hemostasis 1988; 14: 1.

Virchow R. Cellularpathologie in ihrer Begrudung auf physiologische und pathologishce Gewebelehre. Berlin: A. Hirschwald, 1858.

White RH, McGahan JP, Daschbach MM, Harling RP. Diagnosis of deep-vein thrombosis using duplex ultrasound. Ann Intern Med 1989; 111: 297.