Pulmonary embolism

Pulmonary embolism: Sudden lodgment of a blood clot in a pulmonary artery with subsequent obstruction of blood supply to the lung parenchyma.

Etiology and Pathogenesis

The most common type of pulmonary embolus is a thrombus that usually has migrated from a leg or pelvic vein. Most of those that cause serious hemodynamic disturbances form in an iliofemoral vein, either de novo or by propagation from calf vein thrombi. Thromboemboli originate infrequently in the arm veins or in the right cardiac chambers.

Once released into the venous circulation, thromboemboli are distributed to both lungs in about 65% of cases, to the right lung in 25%, and to the left lung in 10%. Lower lobes are involved four times more often than upper lobes. Most thromboemboli lodge in large or intermediate (elastic or muscular) pulmonary arteries; 35% or fewer reach the smaller arteries.

Fat emboli, which can form after fractures, and amniotic fluid emboli are rarer causes. They obstruct primarily the pulmonary microcirculation (arterioles and capillaries rather than pulmonary arteries), which may initiate the adult respiratory distress syndrome (see Ch. 67 ). For discussions of air and gas emboli, see Chs. 129 and 285.

Acute pulmonary embolism (PE) is a dynamic process. Thrombi begin to lyse immediately after reaching the lung. Usually, lysis is complete within several weeks in the absence of preexisting cardiopulmonary disease; in some instances, even large thrombi may lyse in a few days. The physiologic alterations lessen over hours or days as pulmonary circulation improves. However, massive emboli may cause death within minutes or hours, before infarction has time to develop. Infrequently, embolic events recur over a period of months or years, causing progressive pulmonary arterial obstruction with chronic pulmonary hypertension, increasing dyspnea, and cor pulmonale.

The pathogenesis of venous thrombosis is discussed in Ch. 212. The risk is increased in persons who have certain hematologic disorders, those who are immobilized, and those who undergo hip surgery or knee replacement. In many patients, no predisposing factor can be found.

Pathophysiology

The pathophysiologic changes that follow PE involve derangements in pulmonary hemodynamics, gas exchange, and mechanics. The change in cardiopulmonary function is proportional to the extent of obstruction, which varies with the size and number of emboli obstructing the pulmonary arteries, and to the patient's preembolic cardiopulmonary status. The resulting physiologic changes may include pulmonary hypertension with right ventricular failure and shock, dyspnea with tachypnea and hyperventilation, arterial hypoxemia, and pulmonary infarction.

Pulmonary hypertension results from increased pulmonary vascular resistance. Consequently, the right ventricle must generate higher pulmonary arterial pressure to maintain normal cardiac output. Although some degree of pulmonary hypertension can occur after any PE, significant pulmonary hypertension (> 25 mm Hg mean pressure) usually occurs in a previously healthy lung only when > 30 to 50% of the pulmonary arterial tree is occluded. Pulmonary hypertension may be increased by preexisting cardiopulmonary disease (eg, mitral stenosis or COPD). Pulmonary arterial systolic pressure may rise to 100 mm Hg with acute embolism but may reach only 70 to 80 mm Hg if significant tricuspid regurgitation develops. Higher pressures more commonly occur in patients with preexisting cardiopulmonary disease than in those without.

The primary mechanism of increased resistance is obstruction of pulmonary arteries by thrombi, ie, a decrease in the total cross-sectional area of the pulmonary vascular bed. Pulmonary vasoconstriction appears to play a definite but secondary role. Vasoconstriction is partly mediated by hypoxemia, by serotonin release from platelet aggregates on the thrombi, and possibly by other humoral substances, including prostaglandins.

If pulmonary vascular resistance increases acutely to the extent that the right ventricle cannot generate sufficient pressure to maintain cardiac output, hypotension develops and central venous and right atrial mean pressures increase. Cardiogenic shock occurs in persons without preexisting cardiopulmonary disease only after massive PE involving at least 50% and usually 75% or more of the pulmonary vascular bed. With severe hypotension and shock, mean central venous pressure tends to fall.

Tachypnea, often with dyspnea, almost always occurs after a PE. It appears to be due to stimulation of juxtacapillary receptors in the alveolar capillary membrane by swelling of the alveolar interstitial space. This stimulation increases reflex vagal afferent activity, which stimulates medullary respiratory neurons. Consequent alveolar hyperventilation is manifested by a lowered Pa CO2 .

After occlusion of the pulmonary artery, areas of the lung are ventilated but not perfused, resulting in wasted ventilation--the physiologic hallmark of PE--contributing further to the hyperventilatory state.

Depletion of alveolar surfactant within hours after the embolic event results in diminished lung volume and compliance. Reduced lung volume secondary to atelectasis or infarction after PE may be manifest on the chest x-ray by diaphragmatic elevation.

Diminished lung volume and possibly lowered airway PCO 2 may induce bronchoconstriction, leading to expiratory wheezing. Heparin appears to lessen bronchoconstriction, as evidenced by improved maximal expiratory flow rates. Changes in lung mechanics are usually transient, minor, and therefore unlikely to be important in the genesis of prolonged dyspnea. However, they probably contribute to the development of arterial hypoxemia.

Arterial hypoxemia typically occurs with diminished arterial O 2 saturation (Sa O2 <= 94 to 85%), but Sa O2 may be normal. Hypoxemia is due to right-to-left shunting in areas of partial or complete atelectasis not affected by embolization. Characteristically, atelectasis can be partially corrected by deep breathing, either voluntary or induced by a positive pressure ventilator.

Ventilation/perfusion ( / ) imbalance probably also contributes to hypoxemia. The mechanisms responsible for the / imbalance and atelectasis are not fully defined. In massive PE, severe hypoxemia may result from right atrial hypertension that causes right-to-left shunting of blood through a patent foramen ovale. Low venous O 2 tension may also contribute to development of arterial hypoxemia.

Pulmonary infarction (PI) is hemorrhagic consolidation (often followed by necrosis) of lung parenchyma. It does not occur with most pulmonary emboli. When bronchial circulation is intact and normal, PI rarely develops (10% of cases). Collateral circulation from the bronchial artery probably keeps the lung tissue viable despite pulmonary artery blockage. However, patients with previously abnormal pulmonary circulation are prone to develop PI. PI is sometimes due to thrombosis of the pulmonary arteries in situ, as may occur in congenital heart disease associated with severe pulmonary hypertension or in hematologic disorders (eg, sickle cell disease). Infarcts may heal by absorption and fibrosis, leaving a linear scar, or may resorb completely, leaving normal lung tissue (incomplete infarction).

Symptoms and Signs

The clinical manifestations of PE are nonspecific and vary in frequency and intensity, depending on the extent of pulmonary vascular occlusion, preembolic cardiopulmonary function, and the development of PI. Small thromboemboli may be asymptomatic.

The manifestations of PE usually develop abruptly in minutes; those of PI, over a period of hours. They often last several days, depending on the rate of clot lysis and other factors, but usually decrease in intensity daily. In patients with chronic, recurrent small emboli, the symptoms and signs of chronic cor pulmonale tend to develop insidiously over weeks, months, or years.

Embolism without infarction causes breathlessness. Tachypnea is a consistent, often striking feature. Anxiety and restlessness may be prominent.

Pulmonary hypertension, if severe, may cause dull substernal chest discomfort due to pulmonary artery distention or possibly myocardial ischemia. The pulmonary component of the second heart sound may increase, or the aortic and pulmonary components of the second heart sound may widen but with less widening during inspiration. If PE is massive, acute right ventricular dysfunction may result, with distended cervical veins and right ventricular heave and presystolic (S 4 ) or protodiastolic (S 3 ) gallop, sometimes with arterial hypotension and evidence of peripheral vasoconstriction. A significant number of patients may present with light-headedness, syncope, seizures, and neurologic deficits, usually reflecting a transient fall in cardiac output with secondary cerebral ischemia. Cyanosis usually occurs only in patients with massive PE. A small embolus in the periphery of a lung may cause infarction without pulmonary hypertension.

Results of a lung examination are usually normal in the absence of PI. Wheezing is sometimes heard, particularly if underlying bronchopulmonary or cardiac disease is present.

Signs that may indicate PI include cough, hemoptysis, pleuritic chest pain, fever, evidence of pulmonary consolidation or pleural fluid, and possibly a pleural friction rub.

Diagnosis and Differential Diagnosis

The diagnosis of PE with or without PI is often difficult to establish unless special procedures are used; the most important are radioisotope perfusion lung scans and pulmonary arteriography. Differential diagnosis in patients with massive PE includes septic shock, acute MI, and cardiac tamponade. Without infarction, the patient's symptoms and signs may be attributed to anxiety with hyperventilation because of the paucity of objective pulmonary findings. When PI occurs, the differential diagnosis includes pneumonia, atelectasis, heart failure, and pericarditis. A systematic approach to definitive diagnosis is outlined below.

Without infarction, the chest x-ray may be normal, or diminished pulmonary vascular markings in the embolized area may be noted. With infarction, the x-ray frequently shows a peripheral infiltrative lesion, often involving the costophrenic angle, with elevation of the diaphragm and pleural fluid on the affected side. Dilation of the pulmonary arteries in the hilar area, the superior vena cava, or the azygos vein signals pulmonary hypertension and right ventricular strain. Because ECG changes are typically transient, serial tracings are often helpful in diagnosing or excluding acute MI. Changes seen most often with PE include P pulmonale, right bundle branch block, right axis deviation, and supraventricular arrhythmias.

Serum enzyme studies lack sensitivity and specificity and are rarely helpful in diagnosis. The triad of elevated serum LDH and bilirubin and normal AST occurs in < 15% of patients with acute PE and PI. Elevated LDH may occur in as many as 85% of patients with PI but is nonspecific, occurring also in heart failure, shock, pregnancy, kidney or liver disease, anemia, pneumonia, and carcinoma and after surgical procedures. Blood levels of fibrin split products, such as D-dimer, may rise after PE, whether PI occurs or not. However, specificity is low because false-positive results are common and levels are elevated in other conditions, such as the postoperative state. Extreme caution has been recommended in the use of D-dimer tests because data are limited. Some authorities suggest that when clinical suspicion is low, a normal D-dimer may increase the likelihood that no thromboembolic disease is present.

Lung perfusion scans use IV injection of 20- to 50- μ m particles of biodegradable albumin labeled with technetium 99m. These particles ultimately lodge in the small precapillary arterioles of both lungs. Nearly 100% of the particles remain in the lungs, except when right-to-left shunting is present, either at the cardiac or pulmonary level. Regional distribution of these particles is relatively homogeneous in healthy persons but depends on the patient's position and pulmonary blood flow distribution at the time of injection. Visible activity is greatest at the base and gradually diminishes up to the apex, reflecting gravitational effects on perfusion when the patient is sitting. Perfusion deficit, with reduced or absent radioactivity, may result from vascular obstruction, displacement of a lung by fluid, chest masses, any condition causing pulmonary arterial or venous hypertension, or loss of lung parenchyma, as in pulmonary emphysema. Basal perfusion deficits, in which radioactivity is not concentrated at the lung's base, may develop without PE; they may be caused by any process leading to increased pulmonary venous pressure (eg, heart failure, mitral valve disease, or veno-occlusive disease), which can redistribute pulmonary blood flow.

A normal scan excludes life-threatening PE with a high degree of accuracy. Conversely, single or multiple wedge-shaped marginal scan defects, especially in a segmental or lobar distribution, are highly suggestive of vascular obstruction. Acute airway disease, including asthma, or COPD may produce a pattern of focal perfusion deficits, but these deficits are typically accompanied by a corresponding lung ventilation defect not usually found with PE.

When differentiation between PE and COPD is difficult, the xenon-133 lung ventilation scan may be useful. Inhaled radioactive gas distributes with the respiratory air. In acute PE with large perfusion defects, this scan usually shows relatively normal ventilation of affected areas, but with / mismatches. Areas of parenchymal disease (eg, lobar pneumonia) usually have abnormalities of both perfusion and ventilation (a matched defect), with delayed ventilation and trapping of radioactive gas. Matched / defects may also occur with pulmonary edema. Sometimes, matched / defects occur with PE as well, particularly if scanning is performed > 24 h after the event.

Results of scintigraphy are commonly reported as indicating various degrees of probability for PE and must be interpreted cautiously. If lung scans are entirely normal, the diagnosis of PE is virtually ruled out; if they are categorized as highly probable, the positive predictive value approaches 90%. However, although virtually all patients with PE have abnormal scans, < 50% are placed in a high probability category. Clinical assessment helps determine whether pulmonary arteriography is indicated.

Pulmonary arteriography demonstrates emboli and is the most definitive diagnostic test. It should be performed if the diagnosis is uncertain and the need to make the diagnosis with certainty appears urgent. The two primary diagnostic criteria of PE are intra-arterial filling defects and complete obstruction (abrupt cutoff) of pulmonary arterial branches. Other frequent but less conclusive findings include partial obstruction of pulmonary arterial branches with increased caliber proximal to and decreased caliber distal to the narrowing, oligemic zones, and persistence of dye in the proximal portion of the artery during the late (venous) phase of the arteriogram. In lung segments with obstructed arteries, venous filling with contrast medium is delayed or absent.

Additional diagnostic studies to establish the presence or absence of iliofemoral venous thrombotic disease may be useful, particularly when signs of recurrent embolization despite anticoagulant therapy or contraindications to anticoagulant therapy make inferior vena caval interruption (see below) an important therapeutic consideration. For discussions of duplex ultrasonography, plethysmography, and venography, see Diagnosis under Venous Thrombosis in Ch. 212.

Prognosis

Mortality after the initial thromboembolic event varies with the extent of PE and the patient's preexisting cardiopulmonary status. The likelihood that a patient with markedly compromised cardiopulmonary function will die after significant PE is high (probably > 25%). However, a patient with normal cardiopulmonary status is unlikely to die unless the occlusion exceeds 50% of the pulmonary vascular bed. When the initial embolic event is fatal, death often occurs within 1 to 2 h.

The likelihood of a recurrent embolus in an untreated patient is about 50%, and as many as half of these recurrences may be fatal. Anticoagulant therapy reduces the rate of recurrence to about 5%; only about 20% of these will be fatal.

Prophylaxis

In view of the limitations of treatment, prophylaxis is very important. The choice and intensity of prophylactic measures are determined by clinical factors that predispose to venous stasis and thromboembolism (see Table 72-1 ).

Prophylactic regimens for venous thromboembolism include low-dose unfractionated heparin (LDUH), low molecular weight heparin (LMWH), dextran infusion, warfarin, intermittent pneumatic compression (IPC), and graded compression elastic stockings. Aspirin does not help prevent venous thromboembolism in general surgical patients.

Low-dose heparin (LDUH, LMWH) is effective in reducing the incidence of deep vein (calf) thrombosis (DVT) and PE in patients who undergo a variety of elective major surgical procedures. At a blood level of about 1/5 that required to prevent thrombus propagation, heparin activates antithrombin III sufficiently to inhibit factor Xa, which is required to convert prothrombin to thrombin at an early stage in the coagulation sequence. This action prevents the initiation of clot formation but is ineffective once factor Xa has been activated and the process has started.

Both LDUH and LMWH are administered sc, and laboratory monitoring is not required. Although placebo-controlled randomized trials demonstrate no significant increase in major bleeding, the incidence of wound hematomas is increased with both these drugs. LDUH is usually given 2 h preoperatively (5000 U sc) and q 8 or 12 h thereafter for 7 to 10 days or until the patient is fully ambulatory. Of the LMWH preparations, dalteparin (anti-Factor Xa IU) may be given 2500 U once daily, and enoxaparin is usually given 30 mg bid.

Adjusted-dose warfarin is effective in preventing DVT. Warfarin may be given as a fixed low dose of 2 mg/day or as a dose adjusted to mildly prolong the prothrombin time (INR 1.5 to 2.0).

IPC devices provide rhythmic external compression to the legs or legs and thighs. Their effectiveness is approximately equivalent to that of LDUH in reducing the incidence of DVT in general surgery but is inadequate for hip or knee surgery.

Graded elastic stockings reduce the incidence of DVT, but the protective effect on proximal DVT and PE is uncertain. However, combining stockings with other prophylactic measures may give better protection against venous thromboembolism than any one approach alone.

Special considerations regarding prophylaxis are relevant for certain conditions with a high incidence of venous thromboembolism, such as hip fracture and lower extremity orthopedic surgery (see Table 72-2 ). Both LDUH and aspirin are inadequate for hip fracture surgery or hip replacement; LMWH or adjusted-dose warfarin is recommended. For total knee replacement, risk reductions provided by LMWH and IPC are comparable, and the combination should be considered for patients with concomitant clinical risk factors. The regimens for orthopedic surgery may be initiated preoperatively and should be continued for at least 7 to 10 days postoperatively. In selected patients at very high risk for both venous thromboembolism and bleeding, inferior vena caval interruption with placement of a filter is a prophylactic option.

A high incidence of venous thromboembolism is also associated with elective neurosurgery, acute spinal cord injury, and multiple trauma. Although physical methods (IPC, elastic stockings) have been used in neurosurgical patients because of concern about intracranial bleeding, LMWH appears to be an acceptable alternative. The combination of IPC and LMWH may be more effective than either alone in high-risk patients. Limited data support the combination of IPC, elastic stockings, and LMWH in spinal cord injury or multiple trauma. For very high risk patients, inferior vena caval interruption may be necessary.

The most common medical conditions in which prophylaxis is indicated are MI and ischemic stroke. For MI patients, LDUH is effective, and IPC and/or elastic stockings may be used when anticoagulants are contraindicated. LDUH or LMWH can be used in patients with stroke; IPC and/or elastic stockings may be beneficial.

Recommendations for some other medical conditions include LDUH for patients with heart failure; adjusted-dose warfarin (INR 1.3 to 1.9) for those with metastatic breast cancer; and warfarin 1 mg/day for cancer patients with an indwelling central venous catheter.

Treatment

Initial thromboembolic event: Treatment is supportive. Analgesics are given if pleuritic pain is severe. Although anxiety is often prominent, sedatives, especially barbiturates, should be prescribed cautiously. O 2 therapy is indicated when appreciable arterial hypoxemia (Pa O2 < 60 to 65 mm Hg) is present, particularly if cardiac output is reduced. Continuous O 2 should be given, usually by mask or cannula, in a concentration sufficient to raise Pa O2 and Sa O2 to normal levels (85 to 95 mm Hg and 95 to 98%, respectively) or to as near normal as possible (Pa O2 >= 60 mm Hg, Sa O2 > 90%).

In patients with clinical findings suggestive of pulmonary hypertension and acute cor pulmonale, particularly pending diagnostic procedures (eg, lung scanning or arteriography), -adrenergic stimulation may help maintain tissue perfusion because of its effects as a pulmonary vasodilator and cardiotonic. Isoproterenol 2 to 4 mg/L in 5% D/W may be infused at a rate sufficient to maintain systolic BP at 90 to 100 mm Hg under continuous ECG monitoring. Dopamine and norepinephrine have also been used successfully to treat hypotension complicating PE; norepinephrine is preferred when cardiac output is very low. Appropriate drugs may be useful in aborting and preventing supraventricular tachyarrhythmias (see Regular Narrow QRS Tachycardias in Ch. 205). Digitalis should be avoided during acute hypoxemia unless absolutely necessary (eg, for serious arrhythmia or heart failure). When digitalis is given IV, a modest initial dose is usually desirable (digoxin 0.25 to 0.5 mg). Response to therapy in patients suspected of having hemodynamic impairment with acute cor pulmonale may be monitored by serial measurement of arterial blood gases and hemodynamic parameters. A flow-directed balloon (Swan-Ganz) catheter can be used to determine pulmonary artery and wedge pressures, mixed venous blood O 2 saturation and/or content, and cardiac output by the thermodilution technique.

After massive PE: Treatment after massive PE, particularly with hypotension, or after submassive PE with hypotension in patients who have preexisting cardiopulmonary disease may involve thrombolytic therapy or pulmonary embolectomy.

Thrombolytic therapy is now an alternative to embolectomy when massive PE is uncomplicated by hypotension or when systolic BP can be maintained at 90 to 100 mm Hg with moderate dosage of a vasopressor. Streptokinase, urokinase, and tissue plasminogen activator (TPA) enhance the conversion of plasminogen to plasmin, the active fibrinolytic enzyme. Contraindications to thrombolytic therapy include intracranial disease, stroke within 2 mo, active bleeding from any source, preexisting hemorrhagic diathesis (as in impairment of hepatic or renal function), pregnancy, severe or accelerated hypertension (diastolic pressure > 110 mm Hg), and surgery within the preceding 10 days, which is a major limitation to thrombolytic therapy.

If the patient has been receiving heparin, the partial thromboplastin time should be permitted to fall to < 2 times the control before initiating fibrinolytic therapy. Premedication with hydrocortisone sodium succinate 100 mg IV and reinjection q 12 h minimize allergic and pyrogenic reactions to streptokinase. After baseline determination of fibrinogen level or thrombin time, streptokinase 250,000 U is given IV over 30 min, followed by continuous infusion of 100,000 U/h for 24 h. After 3 to 4 h, the fibrinogen level should be about the control, and activated partial thromboplastin time (APTT), thrombin time, or euglobulin lysis time should be prolonged, indicating fibrinolysis. If no change occurs, the patient is resistant to streptokinase and may be given alternative thrombolytic therapy. A priming dose of urokinase 4400 U/kg is given IV over a 10-min period, followed by 4400 U/kg/h for 12 h. Most of the recent experience involves TPA. TPA may be given IV 50 mg/h for 2 h. If repeat pulmonary angiogram shows no evidence of clot lysis and no bleeding complications preclude further therapy, an additional 40 mg may be given over the next 4 h (10 mg/h). After infusion of a thrombolytic drug, the APTT should be allowed to fall to 1.5 to 2.5 times the baseline value before initiation of a sustaining infusion of heparin without a loading dose.

All patients undergoing thrombolytic therapy have an increased risk of bleeding, particularly from recent operative wounds, needle puncture sites, sites of invasive procedures, and the GI tract. Thus, invasive procedures should be avoided. Pressure dressings are usually required to stop oozing; serious or catastrophic bleeding requires stopping the thrombolytic drug and administering cryoprecipitate or fresh frozen plasma. In addition, aminocaproic acid 5 g IV immediately and 1 g/h thereafter for 6 to 8 h or until the bleeding stops may reverse the fibrinolytic state.

Pulmonary embolectomy should be considered when systolic BP is <= 90 mm Hg, urine output is <= 20 mL/h, and Pa O2 is <= 60 mm Hg for as long as 1 h after massive PE. Angiographic confirmation of PE is strongly advised before embolectomy; inferior vena caval interruption and IV heparin therapy generally follow the embolectomy. In the event of cardiac arrest after massive PE, the usual resuscitative measures are ineffective because blood flow through the lungs is obstructed. In this setting, emergency partial (femoral venoarterial) bypass, pending pulmonary embolectomy, may be lifesaving.

Partial interruption of the inferior vena cava by a filter should be considered in certain situations: when anticoagulation is contraindicated, when emboli recur despite adequate anticoagulation, for septic pelvic thrombophlebitis with emboli only when antibiotics and heparin fail, and when pulmonary embolectomy is performed. The filter is placed via catheterization of the internal jugular or femoral veins. The optimum location for the filter is just below the entry of the renal veins. Patients who have had vena caval filter placement may require anticoagulation for at least 6 mo after the procedure for treatment of underlying DVT.

Preventing further thrombus formation and embolization: After initial treatment, prevention becomes the focus. Heparin may be given IV q 4 to 6 h or by continuous IV drip with an infusion pump. However, a hemorrhagic disorder or an active bleeding site is an absolute contraindication to heparin therapy; septic embolization is usually a contraindication. Hemorrhagic complications are reduced by continuous infusion, which obviates the peaks and troughs of blood levels that occur with intermittent injection.

After a rapid IV loading dose of heparin 100 U/kg, heparin is given at a rate to keep the APTT at 1.5 to 2 times control. Achieving a therapeutic APTT in the first 24 h is critical, because failure to do so is associated with a high rate of recurrent venous thromboembolism. APTT may be checked q 4 h after treatment is initiated, and additional bolus dosing may be used to achieve an adequate APTT, followed by adjustment of the infusion rate (see Table 72-3 ). The maintenance dose by continuous infusion is usually 10 to 50 U/kg/h; once a therapeutic level is established, APTT needs to be measured only 1 or 2 times/day.

Oral warfarin sodium may be initiated on the first day of heparin therapy. Warfarin and heparin should overlap for 5 to 7 days, allowing warfarin to take effect, until the INR is in the therapeutic range for two consecutive days. Warfarin sodium 10 mg may be given the first day, with the daily dosage adjusted thereafter to keep the INR at 2.0 to 3.0. The elderly tend to be highly sensitive to warfarin.

The duration of anticoagulant therapy is adjusted individually. In patients with a definable, reversible cause (eg, the postoperative state), anticoagulation may be discontinued after 2 to 3 mo. Otherwise, it may be continued empirically for 3 to 6 mo. A patient with a chronic disorder associated with a high incidence of thromboembolism may need long-term or lifelong anticoagulant therapy.

Complications of anticoagulant therapy: Patients treated with anticoagulants are prone to bleeding complications, some of which may be severe. Periodic platelet counts (in patients taking heparin--see Heparin-Induced Thrombocytopenia in Ch. 133), together with hematocrits and tests for occult blood in stool, are recommended. Patients taking anticoagulants should not be given any drug containing aspirin or other NSAIDs, which can further impair hemostatic mechanisms. Many other drugs, by a variety of mechanisms, can also cause potentially clinically significant drug-drug interactions with oral anticoagulants, enhancing or decreasing their effects. For example, drugs that decrease intestinal synthesis of vitamin K or interfere with other components of normal hemostasis, drugs that interfere with absorption or protein binding, and drugs that increase or decrease hepatic metabolism can modify the pharmacokinetics and pharmacodynamics of warfarin. The direction and extent of effects of these interactions are not completely predictable, but vigilance and more frequent determinations of prothrombin time are indicated when any drug is added or deleted from the regimen of a patient stabilized on an oral anticoagulant. Furthermore, patients should be warned not to take OTC drugs or drugs prescribed by other physicians without first informing their primary care physician.

Other complications of anticoagulant therapy include minor bleeding (ecchymoses at the site of injections, microscopic hematuria, bleeding from gums), which can usually be controlled by withholding the next scheduled dose of heparin and reducing subsequent doses. If major bleeding occurs, protamine sulfate, a protein that combines with heparin to form an inactive complex, should be used to neutralize the anticoagulant effect of heparin. Fifty milligrams (5 mL) diluted with 20 mL of 0.9% sodium chloride solution and injected IV over a 10-min period ( Caution: Rapid injection may cause hypotension, dyspnea, and bradycardia ) neutralizes about 5000 U of heparin and usually suffices to counteract overheparinization. Giving > 100 mg of protamine over a short time is unwise because of its anticoagulant effect. The therapeutic effect of protamine may be evaluated by the APTT. Transfusion therapy may be required for major blood losses but does not reduce the anticoagulant effect of overheparinization. Long-term therapy with heparin leads to osteoporosis and hypoaldosteronism, which causes potassium retention. Uncommon side effects include thrombocytopenia, occasionally with severe thromboembolic shock (see Heparin-Induced Thrombocytopenia in Ch. 133); urticaria; and anaphylactic shock.

As with heparin, the major complication of warfarin therapy is bleeding. Withholding the drug or adjusting the dosage usually controls minor bleeding. For major hemorrhage, 5 to 25 mg (rarely, up to 50 mg) of parenteral vitamin K may be given. In emergency situations of severe hemorrhage, clotting factors may be normalized by giving 200 to 500 mL of fresh whole blood or fresh frozen plasma or by giving factor IX complex parenterally. Purified factor IX preparations should not be used because they do not increase levels of prothrombin, factor VII, or factor X.