Incidence

Thromboembolic disease is one of the most significant potential complications following spine surgery, with rates of acute deep venous thrombosis (DVT) ranging from 0.3% to 31%, and an overall DVT incidence of 2.2%. These rates vary substantially depending on a number of factors. Overall, as expected, the lowest rates occur in younger patients undergoing simple elective procedures, whereas the highest reported rates are those in patients with preexisting risk factors that predispose them to DVT.

Risk Factors

Virchow’s triad of venous stasis, endothelial injury, and hypercoagulability is the classic description of the combination of factors that may predispose a person to DVT. In addition, general clinical risk factors include advanced age, trauma, previous DVT, stroke, malignancy, smoking, and exogenous estrogen replacement. Although numerous systems have been developed to attempt stratification of DVT risk in surgical patients, many are cumbersome and therefore not of practical utility for most surgeons. However, one of the simpler systems involving assignment of a patient into one of four categories on the basis of complexity of procedures (major versus minor), age, and additional risk factors (such as prior DVT, hypercoagulability, and malignancy) serves as a general guide to estimate the venous thromboembolism risk and provide prophylaxis recommendations. In general, spine surgery patients are at a higher risk than general surgery patients secondary to the postoperative immobility that may occur due to preexisting neurologic deficits or postoperative pain. Conversely, the risk is significantly lower than that for patients undergoing lower-extremity surgery such as total hip or knee replacement, which can be associated with DVT rates as high as 50%. Although the overall risk of DVT in spine surgery patients can be described as moderate, the degree of immobility, and thus the risk of DVT, can be directly correlated with the type of spine procedure being performed. For example, patients undergoing a single-level anterior cervical discectomy and fusion are often treated in outpatient surgical centers and discharged home the same day, thereby minimizing the amount of postoperative bed rest and risk of DVT. Patients undergoing more extensive and complex surgery, particularly those with traumatic spinal cord injury (SCI), are at the highest risk. With regard to acute SCI, the incidence of DVT has been reported to range from 10% to 100% without prophylaxis and 0% to 7% with prophylaxis. Interestingly, no correlation has been observed among level of injury, American Spinal Injury Association (ASIA) grade, or spasticity and the incidence of DVT.

Unfortunately, there is a limited amount of evidence regarding the specific risks of DVT in spine surgery patients. One study that used routine venography in patients undergoing spine surgery who did not receive any prophylaxis reported an incidence of 15.5%. It is important to note that none of these patients had any clinical evidence of DVT, underscoring the low sensitivity of physical signs in the diagnosis of DVT. Furthermore, the same authors found that lumbar surgery carries a much higher risk of DVT (21%) as compared to cervical surgery (6%). As previously noted, the risk of DVT may also increase with the complexity of the procedure because more complex operations have longer operative times and often increased postoperative immobility. The use of ventral and lateral approaches further elevates the risk by requiring manipulation of vessels, particularly major veins, and thus increasing the chance for endothelial disruption. Dearborn and colleagues reported an incidence of 6% in patients undergoing combined ventral/dorsal approaches compared to 0.5% in patients in whom only a dorsal approach was employed.

Prevention

Recommendations for DVT prophylaxis for patients undergoing spine procedures are varied and inconsistent. This inconsistency stems largely from the lack of rigorous supporting evidence for many of the prophylactic measures employed. Available prophylactic modalities include the use of gradient compression stockings (GCSs) or intermittent pneumatic compression devices (ICDs), subcutaneous administration of low-dose unfractionated heparin (LDUH) or low-molecular-weight heparin (LMWH), and the placement of an inferior vena cava filter. The first step in determining appropriate DVT prophylaxis for an individual patient is assessing the risk of DVT using the criteria described previously. This risk depends heavily on the procedure being performed and the patient’s age and comorbid conditions. In 2004, the American College of Chest Physicians published guidelines for prevention of venous thromboembolism. These recommendations are summarized in Box 207-1 .

Although these guidelines may serve as a foundation on which clinical decisions may be based, the decision on when and how to appropriately administer DVT prophylaxis for spine surgery patients remains a subject of significant controversy. The risks of each prophylactic intervention must be compared to the risks of DVT in each individual patient. Most spine surgeons routinely prescribe some form of DVT prophylaxis in almost all patients. As previously mentioned, the evidence to support these policies is far from unequivocal. At baseline, most spine patients will be fitted with ICDs or GCSs as a primary method of DVT prophylaxis, as the effectiveness of mechanical prophylaxis has been demonstrated and the risks associated with their use are extremely low. One caveat to the use of ICDs is that their effectiveness depends on perioperative and postoperative employment, as the highest risk for DVT development occurs at induction of anesthesia.

Although the risk of DVT may be further decreased with the use of pharmacologic anticoagulation, the controversy surrounding its use is significant. Several small studies have demonstrated the effectiveness of LDUH and LMWH in spine surgery patients. In a double-blind randomized controlled trial, Agnelli and Becattini demonstrated a reduction in DVT incidence from 30% in patients treated with ICD alone to 17% in patients treated with ICD plus LMWH. Nevertheless, the concerns for increased intraoperative blood loss and postoperative epidural hematoma formation have led to a lack of standardized guidelines for DVT prophylaxis in spine surgery patients.

We adjust the prophylactic regimen according to the neurologic condition, ambulatory status, age, procedure performed, and medical comorbidities of each patient. Preoperatively, hospitalized patients on bed rest or patients with neurologic deficits that affect the lower extremities are treated with ICDs. Intraoperatively, all patients receive continuous treatments with ICDs. It is also customary during prolonged ventral approaches to provide periodic release of retraction to decrease tension of the great vessels. Postoperatively, ICDs alone are continued if the patient will be ambulatory within the first 24 hours. If there are significant neurologic deficits or pain control issues that limit mobility within the first 24 hours, LDUH (given as 5000 IU subcutaneously every 8 hours) is started on the morning of postoperative day 1. If ICDs cannot be tolerated owing to injury of the lower extremities, LDUH is administered perioperatively. For long-term prophylaxis in an outpatient or rehabilitation setting, LMWH is used, the higher cost being traded for improvement of patient compliance and reduction in the need for laboratory monitoring.

Screening and Diagnosis

Clinical diagnosis of DVT remains a concern, as less than 50% of patients will exhibit clinical signs. This raises the question of whether routine screening should be employed as a method of early detection. Whereas some authors advocate routine screening, there does seem to be at least a majority consensus that routine screening for DVT with ultrasound or venography is not clinically indicated after spine surgery. Similarly, a comprehensive review by Furlan and Fehlings concluded that there is insufficient evidence to support routine screening for DVT in patients with acute SCI.

If routine screening is not indicated, the question arises as to the most sensitive and cost-effective method for diagnosis of DVT in patients in whom DVT is suspected. Lower-extremity pain and tenderness, leg edema, and low-grade fevers can be nonspecific indicators of DVT. Even though DVT is confirmed in only 10% to 25% of patients in whom it is suspected clinically, clinical suspicion remains an important first step in the initiation of more accurate diagnostic testing.

Beyond clinical suspicion, objective confirmatory tests remain mandatory for accurate diagnosis of DVT. Contrast venography continues to be the gold standard for diagnosis of DVT against which other tests are measured. No other modality is as sensitive and specific for both proximal and distal DVT. However, high cost, limited availability, patient discomfort, and contrast reactions have led to the increased use of less invasive diagnostic modalities. By using a pressurized cuff, impedance plethysmography measures the change in electrical impedance of the lower extremity in response to occlusion of the deep venous system. The sensitivity and specificity are high for proximal DVT but lower for distal DVT on a single examination. The accuracy can therefore be increased with serial examinations. By comparison, B-mode ultrasonography is as sensitive as plethysmography for proximal DVT and more sensitive for distal DVT. The addition of Doppler flow analysis in conjunction with ultrasonography has demonstrated sensitivity and specificity of 95% to 100% and has therefore become the diagnostic modality of choice in most clinical settings.

Treatment

Management of acute DVT is directed toward reducing both the short-term (pulmonary embolism [PE], clot propagation) and long-term (postphlebitic syndrome) complications. General management includes bed rest, elevation of edematous extremities, and administration of appropriate analgesics (non-platelet-active agents). Definitive management of acute proximal DVT requires a decision regarding the risk of anticoagulation to the patient. If the risk for systemic anticoagulation is acceptable, treatment of established DVT may be initiated in several ways. Because the risks of using oral anticoagulation agents (warfarin) alone have been well documented, a safe and effective treatment strategy must include an initial course of continuous intravenous unfractionated heparin (IVUH), subcutaneous LMWH, or subcutaneous fondaparinux. The need for an initial course of heparin has been demonstrated in a double-blind, randomized trial with a threefold reduction in recurrent venous thromboembolic events compared with oral anticoagulants alone. This is thought to be the result of the long half-life of factor II (compared to proteins C and S), which results in an initial hypercoagulable state at the onset of oral anticoagulant therapy. Recommendations for treatment of acute DVT, summarized in Box 207-2 , are based on the American College of Chest Physician (ACCP) guidelines for patients with DVT.

The aforementioned regimen remains our preferred means of managing venous thromboembolism. It does, however, carry risks of morbidity. The medical and surgical literature contains numerous reports of complications related to heparin therapy. These complications include thrombocytopenia and thrombotic disorders, skin necrosis, priapism, spontaneous hemorrhage, gastrointestinal bleeding, and epidural hematoma formation. Decortication of portions of the vertebral column and the creation of large potential space during exposure predisposes the spine surgery patient to an even higher risk of hemorrhagic complications and hematoma formation. Furthermore, following decompressive surgery, hematomas are often in direct continuity with the thecal sac, placing neural structures at risk of injury and thus necessitating further surgical intervention and its additional risks.

Although it is one of the most commonly prescribed medications for venous thromboembolism and is associated with extensive clinical experience, the use of warfarin has been cumbersome given its narrow therapeutic index necessitating periodical blood testing and multiple drug-drug or food-drug interactions. The new oral anticoagulants (NOACs), on the other hand, have the advantage of rapid onset of action, minimal interactions with drugs and food, wider therapeutic window, and predictable anticoagulation effects. This group of medications includes direct thrombin (factor IIa) inhibitor (e.g., dabigatran), and direct factor Xa inhibitors (e.g., rivaroxaban and apixaban). Studies have shown that treatment with NOACs in patients with acute venous thromboembolism is not inferior to warfarin and the complication of major bleeding is significantly reduced. According to the Randomized Evaluation of Long-Term Anticoagulation Therapy (RE-LY), Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation (ARISTOTLE), and Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared with Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation (ROCKET AF) trials, the risk of intracranial hemorrhage associated with anticoagulation is reduced by 60% to 69% with dabigatran, 33% with rivaroxaban, and 58% with apixaban, when compared to warfarin. Although they have been approved and seem to be alternatives to warfarin, treatment of venous thromboembolism with NOACs following spine surgery has not been extensively studied, there may be some concerns about the lack of specific antidotes, and currently no evidence-based reversal strategies are available if there is a bleeding complication.

Initial Evaluation

Common clinical manifestations of PE include tachypnea, dyspnea, and pleuritic chest pain. The initial evaluation for clinical suspicion of PE includes chest radiograph, arterial blood gas measurements, and electrocardiogram. The arterial blood gas measurement is useful to demonstrate alterations of oxygen transfer that accompany the ventilation of lungs that have a reduction of pulmonary vascular inflow (ventilation/perfusion mismatch). Arterial blood gases typically reveal respiratory alkalosis, a variable reduction in partial arterial oxygen pressure, and widening of the alveolar-arterial oxygen pressure gradient. Chest radiographs and electrocardiograms are more important and are used to rule out other diagnoses, such as pneumonia, pneumothorax, myocardial infarction, or pulmonary edema. Occasionally, the electrocardiogram may reveal right axis deviation or a right bundle branch block that may aid in the diagnosis of PE. Most commonly, chest radiographs reveal nonspecific findings such as pleural effusion, infiltrate, atelectasis, or elevation of the hemidiaphragm or are negative. Measurement of brain natriuretic peptide (BNP) is sensitive but not specific for a diagnosis of PE. Similarly, measurements of serum d -dimer (by enzyme-linked immunosorbent assay) have a high sensitivity but low specificity, particularly in postoperative patients, as the d -dimer may be elevated from the procedure. However, the utility of d -dimer assays remains in that the negative predictive value of levels below threshold (< 500 ng/mL) are sufficient to exclude PE in patients with low or moderate pretest probability.

Diagnostic Modalities

If suspicion remains high for PE after initial evaluation, further diagnostic workup is recommended. The choice of diagnostic tests depends on multiple factors, including clinical probability of PE, availability of modality, patient condition, and cost. Although pulmonary angiography remains the gold standard for diagnosis of PE, its invasive nature and the development of modern imaging techniques have dramatically decreased its use. The widespread availability of spiral computed tomography (CT) scanners has allowed for the increasing use of a CT pulmonary angiogram in the diagnosis of PE. The diagnostic sensitivity and specificity of a CT pulmonary angiogram are 83% and 96%, respectively. In patients who are intolerant to intravenous (IV) contrast or those with prohibitively poor renal function, a ventilation/perfusion scan remains an option for the diagnosis of PE. The diagnostic accuracy of the ventilation/perfusion scan appears to be similar to that of the CT pulmonary angiogram both in overall sensitivity and in the observation that results must be correlated with clinical suspicion. Specifically, a negative scan in a patient with low pretest probability virtually excludes the diagnosis of PE.

Treatment

The treatment of patients with “nonmassive” PE is exactly the same as that previously described for DVT. However, IVUH is recommended over LMWH for the treatment of massive PE and in patients with significant renal impairment. Massive PE with significant hemodynamic compromise requires urgent intervention with acute thrombolysis, surgical embolectomy, or percutaneous transvenous fragmentation or removal of emboli.