Acute Deep Venous Thrombosis

Published on 14/06/2015 by admin

Filed under Surgery

Last modified 14/06/2015

Print this page

rate 1 star rate 2 star rate 3 star rate 4 star rate 5 star
Your rating: none, Average: 0 (0 votes)

This article have been viewed 4446 times

Chapter 49

Acute Deep Venous Thrombosis

Pathophysiology and Natural History

Jordan P. Knepper, Thomas W. Wakefield

The complications of acute venous thromboembolism (VTE), including deep venous thrombosis (DVT), pulmonary embolism (PE), and the postthrombotic syndrome, are the most common preventable cause of hospital death1 and a source of substantial long-term morbidity.24 The impact on health is so great, the Surgeon General of the United States issued a Call to Action to aid in combating VTE.5 An understanding of the epidemiology, pathophysiology, and natural history of VTE is essential in guiding prophylaxis, diagnosis, and treatment. In addition, recognizing underlying risk factors and the multifactorial nature of VTE may aid in the identification of situations likely to provoke thrombosis in both high-risk individuals and those with unexplained thromboembolism. Furthermore, understanding of the natural history of VTE is important to define the relative risks and benefits of anticoagulation as well as the duration of treatment in individual patients.

Intimate to the pathophysiology is the triad initially described by Rudolf Virchow (stasis, endothelial damage or injury, and hypercoagulability). However, an improved understanding of the coagulation and fibrinolytic systems as well as of the role of the vascular endothelium and the inflammatory response in thrombosis and hemostasis has led to the identification of new prethrombotic conditions and also a more thorough appreciation of previously recognized risk factors. It now appears that many venous thrombi arise from the convergence of several risk factors on a background of an imbalance between coagulation and fibrinolysis. Similar factors define the balance between recanalization of the venous lumen and recurrent thrombotic events that may be important determinants of long-term outcome after an episode of acute venous thrombosis.



The incidence of recurrent fatal and nonfatal VTE is estimated to exceed 900,000 cases annually in the United States alone.6 A 35-year population-based study using the Rochester Epidemiology Project database of Olmsted County, Minnesota, demonstrated an overall average age- and sex-adjusted annual VTE incidence of 122 per 100,000 person-years (DVT, 56 per 100,000; PE, 66 per 100,000).7 This study also demonstrated higher age-adjusted rates among men than among women (134 vs 115 per 100,000 respectively). First-time VTE is approximated to occur in about 250,000 U.S. white individuals annually.8 Compared with other racial populations, whites have a lower incidence of VTE than African Americans (104 vs 141 per 100,000) and a higher incidence of VTE than Hispanics and Asian/Pacific Islanders combined (104 vs 21 per 100,000).9 However, the problem of VTE is not just isolated to the United States; it is a global issue. The estimates of VTE across the European Union were 684,019 cases of DVT and 434,723 cases of PE, with 543,454 VTE-related fatalities.10

Populations Affected

Although estimated incidences of VTE in the community and hospital have been well published, those individuals who are at risk for DVT are less well studied. One large epidemiologic study addressed this issue using the Seventh American College of Chest Physicians Conference on Antithrombotic and Thrombolytic Therapy guidelines for VTE prevention.11 Of the estimated 38 million patient discharges in 2003, 31% were considered to be at risk for VTE because of a major surgery or a medical illness.12

Similarly, the incidence of VTE varies with the population studied, the use of thromboprophylaxis, the intensity of screening, and the accuracy of the diagnostic test employed. For example, individuals with acute spinal cord injury who were screened systematically with venography demonstrated DVT at a rate of 81%.13 However, medical-surgical intensive care unit (ICU) patients who received thromboprophylaxis had a DVT rate reported at 10% to 18% compared with a rate of 25% to 32% for those who were not given DVT prophylaxis.14 Interestingly, the risk of VTE in the critically ill is not limited to the time actually spent in the ICU. A single-center study noted that of the VTEs diagnosed in the critically ill, 64% were diagnosed with a VTE after discharge from the ICU; prolonged immobility after discharge from the ICU may have contributed to the high rate of DVT.15 Similarly, prolonged immobility contributes to increased rates of VTE in another population of patients, nursing home residents.16 In summary, it appears that medical-surgical ICU patients are at lower risk for DVT compared with acute spinal cord injury, trauma, or neurosurgery patients; at a comparable risk to that of patients who have had major orthopedic surgery; and at higher risk than that of medical-surgical ward patients.14,17 Furthermore, a more recent study noted a high (15.2%) rate of DVT in critically ill trauma patients within the first week that did not vary regardless of whether prophylaxis was used.18


VTE is a devastating acute disease process, but it may be manifested chronically. Nearly 30% of patients have a recurrence in a 10-year time span.19 Recurrences are more likely with the same event type as that of the incident event, that is to say, those who initially developed a PE are more likely to develop another PE instead of a DVT.20


The severity of PE is shown in 30-year autopsy studies, which demonstrated a 26% incidence of PE in hospitalized patients, of which 9% were fatal. This translates to a 1% incidence of PE and a 0.36% incidence of death from PE in all hospitalized patients per year.21

Risk Factors

DVT occurring in the setting of a recognized risk factor is often defined as a secondary event; in the absence of risk factors, it is termed primary or idiopathic.22 Known risk factors for DVT are listed in Table 49-1.

The high incidence of acute DVT in hospitalized patients, the availability of objective diagnostic tests, and the existence of clinical trials evaluating prophylactic measures have helped to identify high-risk groups in this population. In contrast, the risk factors for acute DVT are less well defined in population-based studies. Nonetheless, substantive differences have been noted in the distribution of risk factors between inpatients and outpatients.23 Malignant disease, surgery, and trauma within the previous 3 months remain significant risk factors for outpatient thrombosis, whereas the frequency of surgery and malignant disease is higher among inpatients with DVT.24,25 Approximately 47% of outpatients with a documented DVT have one or more recognized risk factors,24 and the incidence of VTE increases with the number of risk factors.26 Although some investigators have found the risk of acute DVT to be significantly higher only for those with three or more risk factors,25 others have reported the relative risk to increase from 2.4 in those with one risk factor to more than 20 in patients who have three or more risk factors.27 Heit et al28,29 identified several risk factors for VTE, including increasing age, male gender, surgery, trauma, hospital or nursing home confinement, malignant disease, neurologic disease with extremity paresis, central venous catheter or transvenous pacemaker, prior superficial vein thrombosis, and varicose veins. Among women, additional risk factors include pregnancy, oral contraceptives, and hormone replacement therapy. Independent predictors for recurrent DVT, which occurs in 30% of people within 10 years, include increasing age, obesity, malignant neoplasm, and extremity paresis.


VTE occurs in all ages, although a higher incidence has consistently been associated with advanced age. Some investigators have identified cutoff points at which the risk of DVT is raised, whereas others have reported a relative increase with advancing age. In a community-based study of phlebographically documented DVT, the yearly incidence of DVT was noted to increase progressively from almost 0 in childhood to 7.65 cases per 1000 in men and 8.22 cases per 1000 in women older than 80 years.30 The incidence of DVT increased 30-fold from those aged 30 years to those older than 80 years. Rosendaal27 similarly noted an incidence of 0.006 per 1000 children younger than 14 years, which rose to 0.7 per 1000 among adults 40 to 54 years old. Furthermore, Hansson et al31 found the prevalence of objectively documented thromboembolic events among men to increase from 0.5% at the age of 50 years to 3.8% at the age of 80 years.

The influence of age on the incidence of VTE is likely to be multifactorial. The number of thrombotic risk factors increases with age, three or more risk factors being present in only 3% of hospitalized patients younger than 40 years but in 30% of those 40 years and older.26 Interestingly, it also appears that the number of risk factors required to precipitate thrombosis decreases with age.27 This may be related to an acquired prethrombotic state associated with aging as higher levels of thrombin activation markers are found among older people.32 Advanced age also has been associated with anatomic changes in the soleal veins and more pronounced stasis in the venous valve pockets.33,34 We have also noted biologic changes with aging in our research laboratory, such as elevations in P-selectin, tissue factor, and procoagulant microparticles, supporting the effect of age on venous thrombogenesis.

Venous diseases, including VTE, are usually regarded as rare in young children,3 with an incidence of 0.006 per 1000 children younger than 14 years,27 whereas the incidence in hospitalized children younger than 18 years has been estimated to be 0.05%.35 Early mobilization and discharge may partially explain the lower incidence in children.35 However, the diagnosis is often not considered in pediatric patients, and few studies have systematically evaluated children for DVT. Over time, the number of recognized cases in children hospitalized has increased from 0.3 to 28.8 per 10,000 from 1992 to 2005.36

VTE in children is almost always associated with recognized thrombotic risk factors,27,3739 and multiple risk factors are often required to precipitate thrombosis.27 Venous thrombosis is more common in those children hospitalized in the ICU, those with spinal cord injuries, and those undergoing prolonged orthopedic immobilization, although the incidence is substantially lower than in corresponding adult populations. DVT may occur in as many as 3.7% of pediatric patients immobilized in halo-femoral traction for preoperative treatment of scoliosis,40 4% of children hospitalized in the ICU,41 and 10% of children with spinal cord injuries.42,43 Symptomatic postoperative DVT is regarded as unusual in children, although there are few data from studies employing routine surveillance,44 and autopsy-identified PE is approximately four times more frequent in pediatric patients who have undergone surgery than in the general pediatric medical population.37 Other thrombotic risk factors in hospitalized children are local infection and trauma, immobilization,39 inherited hypercoagulable states,27 oral contraceptive use,45 lower limb paresis,40 and use of femoral venous catheters.46 In addition, the risk factors for inpatient DVT in children include catheters and concomitant severe respiratory, oncologic, and infectious diseases; outpatient DVT is often associated with a prior DVT and thrombophilia.36


Many of the clinical presentations of thromboembolism support immobilization as a risk factor for VTE. In addition, stasis in the soleal veins and behind the valve cusps is exacerbated by advancing age and inactivity of the calf muscle pump,34 both of which are associated with an increased risk of DVT. The prevalence of lower extremity DVT in autopsy studies also parallels the duration of bed rest, with an increase during the first 3 days of confinement and a rapid rise to high levels after 2 weeks. DVT was found in 15% of patients dying after 0 to 7 days of bed rest, in comparison with 79% to 94% of those dying after 2 to 12 weeks.33 Preoperative immobilization is also associated with a twofold increase in risk of postoperative DVT,47 and DVT among stroke patients is significantly more common in paralyzed or paretic extremities (53% of limbs) than in nonparalyzed limbs (7%).48 Patients with neurologic disease and extremity paresis or paralysis have a threefold higher risk for DVT and PE that appears to be independent of hospital confinement.28


Immobilization as a thrombotic risk factor extends to include prolonged travel, particularly the “economy class syndrome,” which occurs in people who have sat in a cramped position during extended aircraft flights.49 Several case series have reported the occurrence of PE in relation to extended travel,4954 although none has rigorously examined the prevalence relative to that of the general population, and few have thoroughly reported the presence of other risk factors. A high prevalence of preexisting venous disease and other thrombotic risk factors in this group of patients has sometimes been noted.32,54 The question of prolonged travel as a risk factor is moderated by observations that extreme duration of venous stasis alone may fail to produce thrombosis55 and that no consistent rheologic or prothrombotic changes have been demonstrated during prolonged travel.32,56 However, PE is the second leading cause of travel-related death, accounting for 18% of 61 deaths, suggesting that a relationship cannot be excluded.53 After a consensus meeting, the World Health Organization published the following conclusions57:

More evidence of a connection between travel and DVT and PE has accumulated. In a case-control study, Ferrari et al58 found that long-distance travel increased the risk of DVT with an odds ratio (OR) of 4.0, and Samama59 made similar observations (OR, 2.3). Scurr et al60 found a 10% risk of calf DVT in patients who traveled without compression stockings. Lapostolle et al61 observed that during an 86-month period, 56 of 135.3 million airline passengers had severe PE. The frequency among those who traveled more than 5000 km was 150 times as high as the frequency among those who traveled less than 5000 km. In another case-control study, Paganin et al62 observed a high incidence of VTE in patients with risk factors for DVT who traveled long distances. In particular, history of previous VTE (OR, 63.3), recent trauma (OR, 13.6), presence of varicose veins (OR, 10), obesity (OR, 9.6), immobility during flight (OR, 9.3), and cardiac disease (OR, 8.9) increased the risk of DVT. These investigators concluded that low mobility during flight was a striking modifiable risk factor for development of PE and that travelers with risk factors should increase their mobility.63

Hughes et al64 noted that the frequency of VTE is increased even in travelers otherwise at low to moderate risk. In a prospective study that involved 878 long-distance air travelers, the frequency of VTE was 1.0% (9 of 878; 95% confidence interval, 0.5-1.9), which included 4 cases of PE and 5 of DVT. Two of the patients traveled in business class, five used aspirin, and four wore compression stockings.

History of Venous Thromboembolism

Approximately 23% to 26% of patients presenting with acute DVT have a previous history of thrombosis,23,30 and histologic studies confirm that acute thrombi are often associated with fibrous remnants of previous thrombi in the same or nearby veins.65 Depending on sex and age, population-based studies have demonstrated that recurrent thromboembolism occurs in 2% to 9% of people with a previous episode of thromboembolism.3

The risk of recurrent thromboembolism is higher among patients with idiopathic DVT.66 In addition, primary hypercoagulability appears to have a significant role in many recurrences. Simioni et al66 reported the cumulative incidence of recurrent thrombosis among patients who are heterozygous for the factor V Leiden mutation to be 40% at 8 years of follow-up, 2.4-fold higher than in patients without the mutation, although the importance of heterozygous factor V Leiden to recurrent thrombosis has been questioned.67 den Heijer et al68 estimated that 17% of recurrent thromboembolic events may be due to hyperhomocysteinemia. A similar relationship between impaired fibrinolysis and recurrent DVT has been suggested by several investigators, although the methodologic validity of these findings has been questioned.69

Malignant Disease

Approximately 20% of all first-time VTE events are associated with malignant disease.70 An estimated 1 in 200 individuals with malignant disease will develop either DVT or PE, a fourfold higher risk than in those without malignant disease.28 Considering all-cause mortality of in-hospital death for cancer patients, one in seven will die of PE.71 Unfortunately, the discovery of an occult malignant neoplasm associated with an otherwise first-time idiopathic VTE is as high as 12% to 17% in some series.72,73 In another 5% to 11% of patients, malignant disease appears within 1 to 2 years of presentation for DVT.7274 Related to this, several series have documented a significantly higher risk of malignant disease in patients with idiopathic DVT.72,75 Among such patients, 7.6% have been noted to have a malignant neoplasm during follow-up, with an odds ratio of 2.3 in comparison with those with secondary thrombosis.73 The incidence of occult malignant disease diagnosed within 6 to 12 months of an idiopathic DVT is 2.2 to 5.3 times higher than that expected from general population estimates.76,77 Furthermore, in individuals with recurrent idiopathic DVT, a malignant neoplasm can be identified 17% of the time.73 A correlation with location of tumor reveals that the highest rates of VTE are asso­ciated with pancreatic malignant neoplasms, followed by kidney, ovary, lung, and stomach malignant neoplasms.78

The underlying mechanisms contributing to the hypercoagulable state in malignant disease have been well studied and are multifactorial. Venous compression secondary to tumor growth, cancer-associated thrombocytosis, immobility, indwelling central lines, and chemotherapy or radiation therapy are risk factors that increase the possibility of VTE.79 However, the systemic prothrombotic response seen in malignant disease is mediated by cytokines, inhibitors of fibrinolysis, and procoagulants.80 Tumor cells can directly initiate hemostasis through constitutive expression of tissue factor. Tissue factor is not normally expressed on resting vascular endothelium but rather induced by chemical mediators during times of inflammation or vessel damage to bind factors VII and VIIa. This complex of tissue factor and factor VII activates factors X and XI through proteolysis, leading to the generation of thrombin.81 Another mediator in VTE associated with malignant disease is cancer procoagulant. The role of cancer procoagulant in coagulation is limited to its association with malignant disease as it has not been identified in normal healthy tissue. Cancer procoagulant serves as a direct activator of factor X independent of the presence of factor VIIa. Although cancer procoagulant has not been studied as intensely as tissue factor, the characterization of cancer procoagulant and hypercoagulability in acute myelogenous leukemia has been well published.82,83 The platelet adhesion molecules glycoprotein Ib and glycoprotein IIb/IIIa have also been identified on tumor cells. Such expression allows platelet activation and aggregation.84

The prothrombotic cytokines, such as vascular endothelial growth factor, tumor necrosis factor-α, and interleukin-1, mediate their actions through induction of tissue factor on vascular endothelium, monocytes, and leukocytes.81,82 In addition, interleukin-1 and tumor necrosis factor-α downregulate the expression of thrombomodulin, the receptor for thrombin, on the endothelial surface. This results in a decrease in thrombin-thrombomodulin complexes, the activating complex of protein C, a natural anticoagulant.81,82 Furthermore, interleukin-1 and tumor necrosis factor-α stimulate the production of plasminogen activator inhibitor-1 (PAI-1), the main physiologic inhibitor of fibrinolysis.85

As many as 90% of patients with cancer have abnormal coagulation parameters, including increased levels of coagulation factors, elevated fibrinogen or fibrin degradation products, and thrombocytosis.86 Elevated fibrinogen and thrombocytosis are the most common abnormalities, perhaps reflecting an overcompensated form of intravascular coagulation.86 Levels of the coagulation inhibitors antithrombin, protein C, and protein S also may be reduced in malignant disease.87

Markers of activated coagulation are elevated in the majority of patients with solid tumors and leukemia.86 Fibrinopeptide A levels reflect tumor activity, decreasing or increasing in response to treatment or progression of disease, suggesting that tumor growth and thrombin generation are intimately related.86 Furthermore, these levels may fail to normalize after administration of heparin to patients with cancer and DVT, perhaps explaining why these DVTs may be refractory to anticoagulants.86

VTE is also associated with the treatment of some cancers. DVT complicates 29% of general surgical procedures for malignant disease.87 Preoperative activation of the coagulation system, as reflected by elevated thrombin-antithrombin complex values, is associated with a 7.5-fold increased risk of postoperative DVT.86 Some chemotherapeutic regimens also predispose to DVT, and thrombotic complications may be as common as the more widely recognized infectious complications.88 VTE has been reported in up to 6% of patients undergoing treatment for non-Hodgkin’s lymphoma and in 17.5% of those receiving therapy for breast cancer.89,90 Among patients with stage II breast cancer, thrombosis was significantly more common in those randomly assigned to 36 weeks of chemotherapy (8.8%) than in those receiving only 12 weeks of treatment (4.9%).88 Potential thrombogenic mechanisms associated with chemotherapy include direct endothelial toxicity, induction of a hypercoagulable state, reduced fibrinolytic activity, tumor cell lysis, and use of central venous catheters.89,90 Some intravenous chemotherapeutic agents are associated with activation of coagulation and increased markers of thrombin generation, a response that is blocked by pretreatment with heparin.91


The high incidence of postoperative DVT as well as the availability of easily repeatable, noninvasive diagnostic tests has allowed a greater understanding of the risk factors associated with surgery than in other conditions. Surgery constitutes a spectrum of risk that is influenced by age of the patient, coexistent thrombotic risk factors, type of procedure, extent of surgical trauma, length of procedure, and duration of postoperative immobilization.92 The type of surgical procedure is particularly important.47 The overall incidence of DVT is approximately 19% in patients undergoing general surgical operations and 24% in those having elective neurosurgical procedures; among those undergoing surgery for hip fracture, hip arthroplasty, and knee arthroplasty, it is 48%, 51%, and 61%, respectively.92 On the basis of these data, patients can be classified as being at low, moderate, or high risk for thromboembolic complications, as shown in Table 49-2.

Approximately half of postoperative lower extremity thrombi develop in the operating room; the remainder occur during the next 3 to 5 days.93 However, the risk for development of DVT does not end uniformly at hospital discharge. In one study, 51% of the thromboembolic events occurred after discharge from gynecologic surgical procedures.94 Similarly, up to 25% of patients undergoing abdominal surgery have DVT within 6 weeks of discharge.95 Heit et al94 found a nearly 22-fold higher risk of DVT and PE among patients who were hospitalized after previous surgery.

All components of Virchow’s triad may be present in the surgical patient—perioperative immobilization, transient changes in coagulation and fibrinolysis, and potential for gross venous injury. Immobilization is associated with a reduction in venous outflow and capacitance during the early postoperative period.96 Surgery is also accompanied by a transient, low-level hypercoagulable state, presumably mediated by the release of tissue factor, which is marked by a rise in thrombin activation markers shortly after the procedure begins.96 The thrombogenic potential of different surgical procedures appears to differ, with greater rises in thrombin activation markers during hip arthroplasty than after laparotomy.97 Increased levels of PAI-1 are also associated with a decrease in fibrinolytic activity on the first postoperative day, the “postoperative fibrinolytic shutdown.”98 This relationship between impaired fibrinolysis and postoperative DVT may be particularly important69; preoperative and early postoperative elevations in PAI-1 correlate with the development of thrombosis in orthopedic patients.98


The prevalence of DVT among autopsied trauma casualties has been reported to be as high as 65%, comparable to the 58% incidence among injured patients in modern venographic series.13 Substantially lower DVT rates, ranging from 4% to 20%,99 have been noted in series employing duplex ultrasonography, although many patients studied were receiving prophylaxis and the limitations of ultrasound in screening of asymptomatic patients are well recognized. Recent trauma was associated with a nearly 13-fold increase in risk.94

Although the risk of DVT may be less than 20% in some injured patients,13 certain subgroups are at particularly high risk. Age (OR, 1.05 for each 1-year increment), blood transfusion (OR, 1.74), surgery (OR, 2.30), fracture of the femur or tibia (OR, 4.82), and spinal cord injury (OR, 8.59) have been significantly associated with the development of DVT in this population.13 Other reported risk factors are a hospital stay longer than 7 days,100 increased Injury Severity Score,101,102 Trauma Injury Severity Score of 85 or less,100 pelvic fractures,103 major venous injury,104 presence of femoral venous lines,105 duration of immobilization,106 and prolongation of the partial thromboplastin time.107

As with postoperative DVT, several pathophysiologic elements may be responsible for the high incidence of DVT in trauma patients. Immobilization by skeletal fixation, paralysis, and critical illness are associated with venous stasis, whereas mechanical injury is important after direct venous trauma and central venous cannulation. Less well appreciated is the hypercoagulable state after depletion of coagulation inhibitors and components of the fibrinolytic system. Fibrinopeptide A levels rise after injury,108 consistent with activation of coagulation, whereas fibrinolytic activity has been found to increase initially and then to decrease.109,110

Inherited Thrombophilia

Prothrombin Gene Mutation.

The prothrombin G20210A gene mutation was demonstrated when 28 families with a documented history of VTE were noted to have an 18% incidence of a transition from guanine to adenine at the 20210 nucleotide on the prothrombin gene. A comparison to healthy controls revealed only a 1% incidence of the same mutation.111 In this same study, Poort also reported that in 87% of individuals with such a mutation, the prothrombin levels were greater than 115% of normal. Thus, elevated levels of prothrombin, the precursor for thrombin, are associated with an increased risk of VTE. A large multicenter study revealed the incidence of this mutation in the general European population to be 0.7% to 4%.112 This mutation is rare in individuals of Asian or African descent. Additional investigations have shown that in those with a spontaneous DVT, the incidence of the mutation may range from 7% to 16%, depending on the patient population studied.113 The risk of VTE in the presence of the prothrombin G20210A mutation is threefold higher in heterozygotes than in those with the wild type, and the presence of homozygosity further increases this risk. The prothrombin G20210A gene mutation is frequently coinherited with factor V Leiden mutation in approximately 1% to 10% of symptomatic carriers.114,115

Factor V Leiden Mutation.

Resistance to activated protein C, characterized by the failure of exogenous activated protein C to prolong the activated partial thromboplastin time, was initially described by Dahlbäck in 1993.116 Subsequent studies from the Leiden University hospital in the Netherlands revealed this laboratory phenotype to be the result of a mutation in the gene for factor V of the coagulation cascade in 95% of instances. The transition mutation causes replacement of arginine with glutamine at position 506, such that factor Va is resistant to cleavage by activated protein C and is therefore inactivated at a decreased rate.117 Thus, more factor Va is available to participate in the generation of thrombin. In addition, after cleavage by activated protein C, the degradation product of factor V in combination with protein S acts as a cofactor for activated protein C, increasing the degradation of factor VIIIa as well as additional factor Va.118 The factor V Leiden mutation is inherited in an autosomal dominant manner and is 10 times more common than other heritable defects.119 The factor V Leiden mutation is present in 5% to 8% of European whites, with prevalence of up to 15% in some areas of Greece, Sweden, and Lebanon.120,121 In a study of 4047 Americans, the carrier frequencies of the mutation were 5.3% for whites, 2.2% for Hispanic Americans, 1.25% for Native Americans, 1.2% for African Americans, and 0.45% for Asian Americans. Carrier frequencies were similar in white men and women in this study.122 Factor V Leiden mutation has been reported in 20% to 50% of first-time VTE events; thus it is the most common cause of inherited thrombophilias.123 The relative risk of DVT is increased sevenfold in those who are heterozygous for the mutation. However, in individuals who are homozygous for the mutation, the relative risk is increased 80-fold. In addition, homozygous patients develop DVT at a younger age than do those who are heterozygous.124

The etiology of activated protein C resistance is not limited to factor V Leiden mutations. Increased plasma levels of factor VIII, the presence of antiphospholipid antibodies, older age, pregnancy, and the use of estrogens are well-known causes of increased activated protein C resistance.125

Protein C.

Protein C is a vitamin K–dependent serine protease that inhibits the coagulation system by inactivating factors Va and VIIIa, the two cofactors necessary for activation of factor X and thrombin generation.126 Once it is activated by the thrombin-thrombomodulin complex, activated protein C can be enhanced by protein S.127,128 Either an acquired or inherited reduction of protein C activity increases thrombotic risk. Heterozygous protein C deficiency is inherited in an autosomal dominant fashion; this deficiency may be classified as type I, characterized by a reduction in both antigenic and functional levels, or as type II, in which antigen levels are normal and functional activity is decreased.126 In a study of 10,000 healthy blood donors, the observed prevalence of inherited protein C deficiency was 1.45 per 1000.129 Furthermore, a study of 2132 patients with VTE found that 3.2% had protein C deficiency.130 It is estimated that individuals who are heterozygous for protein C deficiency have a sevenfold higher risk for development of VTE over baseline.131 Other than VTE, protein C deficiency is also associated with neonatal purpura fulminans and warfarin skin necrosis.

Neonatal purpura fulminans can result from either a homozygous or a double heterozygous genotype. It is characterized by microvascular thrombosis in the dermis, leading to severe ecchymoses followed by perivascular hemorrhage, extensive venous and arterial thrombosis, necrosis, laboratory findings consistent with disseminated intravascular coagulation, and extremely low levels of protein C.132 Warfarin-induced skin necrosis may occur in individuals with protein C deficiency during the first several days of warfarin therapy. Warfarin initiation can reduce protein C anticoagulant activity to 50% within 1 day.133 With the exception of factor VII, other procoagulant vitamin K–dependant clotting factors decline at a slower rate, inducing a transient hypercoagulable state. The clinical manifestations of warfarin-induced skin necrosis are similar to those seen in neonatal purpura fulminans.134 Acquired deficiency of protein C may also be seen in acute thrombotic events, liver disease, renal disease, disseminated intravascular coagulation, hemolytic-uremic syndrome, chemotherapy, thrombotic thrombocytopenic purpura, and acute infections.135

Protein S.

Protein S is a vitamin K–dependent cofactor for the protein C–mediated inactivation of factors Va and VIIIa. In addition, protein S can directly inhibit the clotting pathway through interactions with factors Va and Xa. Protein S is synthesized by hepatocytes and megakaryocytes and circulates bound to C4b-binding protein or as a free form; only the free form is able to interact with protein C.136 Protein S deficiency is autosomal dominant and is more common than protein C deficiency. The incidence of protein S–associated VTE is 5% to 7%; the prevalence in the general population is 0.13%.137138 It is estimated that individuals who are heterozygous for protein S deficiency have an 8.5-fold higher relative risk for development of VTE.139

There are more than 130 types of protein S mutations, which can fall into three general categories. Type I deficiency is quantitative, type II deficiency is qualitative, and type III deficiency is caused by mutations that increase the affinity of protein S for C4b-binding protein.125 The clinical presentation of patients with heterozygous protein S deficiency is similar to that of patients with protein C deficiency: an increased incidence of VTE and warfarin-induced skin necrosis.140 Homozygous individuals are rare but will present with neonatal purpura fulminans. Acquired protein S deficiency has been reported in disseminated intravascular coagulation, diabetes mellitus, pregnancy, oral contraceptive use, nephritic syndrome, liver disease, and essential thrombocythemia.126


Antithrombin, a vitamin K–independent glycoprotein, inhibits thrombin as well as factors Xa, IXa, XIa, and XIIa. The action of antithrombin is accelerated by heparin.141 Antithrombin deficiency is inherited in an autosomal dominant manner. The majority of patients are heterozygous; homozygotes usually die in utero. Type I deficiency is a quantitative defect, whereas type II is a qualitative deficiency.125 The incidence of antithrombin deficiency–associated VTE has been reported at 0.5% to 3%, with a prevalence in the general population of 0.2%.130,141 However, it is estimated that heterozygotes for this deficiency have a 20-fold higher risk for development of VTE over that of the general population.142 The clinical presentations include thrombosis of the iliofemoral veins, upper limb deep veins, mesenteric veins, vena cava, renal veins, and retinal veins as well as cerebral venous thrombosis and Budd-Chiari syndrome.125,143


High plasma levels of homocysteine have been associated with an increased risk for both venous and arterial thrombosis.144 The metabolic breakdown of homocysteine involves N5-methyltetrahydrofolate reductase (MTHFR). Individuals who are homozygous for a substitution of cytosine by thymine on the MTHFR gene at position 677 (C677T) and have inadequate dietary intake may achieve moderately increased homocysteine levels.145 Despite this association, data supporting the MTHFR C677T mutation as an independent risk factor for VTE are lacking, but a new study is helping to provide mounting evidence for the importance of hyperhomocysteinemia in VTE.146

Other Inherited Thrombophilias.

Other inherited thrombophilias, such as dysfibrinogenemia and elevated clotting factors, are less well studied. Inherited defects of fibrinogen are detected in less than 1% of those presenting with first-time VTE.147 In such defects, fibrinogen may be unable to bind thrombin, or the fibrin itself is unable to be lysed by plasmin. Clinically, the majority of these patients are asymptomatic, yet 25% may present with bleeding and an additional 20% with thrombosis.148 Elevated levels of factors VIII, IX, and XI have been demonstrated to be an independent risk factor for VTE.149151 Individuals with elevations of factors IX and XI above the 90th percentile in healthy controls have an odds ratio of 2.3 and 2.2, respectively, for DVT compared with those with levels below the 90th percentile.149,150 Elevated factor VIII levels have been reported in up to 25% of individuals presenting with VTE.152

Others have noted residual risk for thrombosis not explained by known inherited thrombophilic disorders. This additional risk is indirect evidence that more inherited thrombophilias are yet to be identified.153


The incidence of VTE in the pregnant population is 6 to 10 times greater than in matched controls,154 leading to approximately 10% of all maternal deaths.155 By clinical evaluation, VTE has been reported at a rate of 1.3% to 7% during pregnancy and 6.1% to 23% in the postpartum period.156 However, studies that employed venography, Doppler ultrasonography, or ventilation-perfusion scans for evaluation of clinically suspected thromboembolism have suggested an incidence of 0.029% to 0.055% in this population.157 The risk of thrombosis appears to be two to three times greater during the puerperium, with the highest incidence found after cesarean section.158 Interestingly, when VTE has been objectively documented, the occurrence of thrombosis is equally distributed throughout all trimesters.159

DVT in pregnancy has been attributed to impaired venous outflow due to uterine compression, and 97% of reported thromboses have been isolated to the left leg.159 Furthermore, pregnancy is associated with a transient hypercoagulable state due to increases in the levels of fibrinogen, von Willebrand factor, and factors II, VII, VIII, and X. Compounding this, acquired functional resistance to activated protein C is also seen during pregnancy.159,160 Similarly, protein S levels are decreased by 50% to 60% early during pregnancy, with free protein S levels comparable to those in hereditary heterozygous protein S deficiency.161,162 The fibrinolytic system is also altered in pregnancy; levels of tissue plasminogen activator are decreased, and plasminogen activator inhibitors 1 and 2 are increased.163

Both retrospective and prospective studies have demonstrated that between 30% and 50% of women with a pregnancy-associated VTE also have an inherited thrombophilia. This high incidence has led to the recommendation of screening for thrombophilia in those pregnant patients with a personal or family history of VTE.164,165 The risk of puerperal DVT also increases with maternal age, suppression of lactation, hypertension, and assisted delivery but not with the number of pregnancies.166

Oral Contraceptives and Hormonal Therapy

As suggested by case reports in the early 1960s, further studies have now established the use of oral contraceptives as an independent risk factor for the development of DVT. These studies noted odds ratios of 3.8 to 11.0 for thrombosis167; an unweighted summary relative risk among 18 controlled studies was 2.9.168 Approximately one quarter of idiopathic thromboembolic events among women of childbearing age have been attributed to oral contraceptives. The risk of hospital admission for a thromboembolic event, including cerebral thrombosis, has been estimated to be 0.4 to 0.6 per 1000 for oral contraceptive users, compared with 0.05 to 0.06 for nonusers.169 Early studies also suggested that thromboembolism is responsible for approximately 2% of deaths in young women, with contraceptive-associated mortality rates of 1.3 and 3.4 per 100,000 among women aged 20 to 34 years and 35 to 44 years, respectively.170 The increased risk of thromboembolism appears to diminish soon after oral contraceptives are discontinued and is independent of the duration of use.171

Risk is correspondingly higher when oral contraceptive use is combined with other factors, such as surgery172 and inherited inhibitor deficiencies.173 The factor V Leiden mutation may be particularly important in this regard; resistance to activated protein C has been reported in up to 30% of patients with contraceptive-associated thromboembolism.174 The use of third-generation oral contraceptives may act synergistically with the factor V Leiden mutation, raising thromboembolic risk 30-fold to 50-fold.175177

The risk of VTE among users of third-generation oral contraceptives may be up to eight times that of young women who do not use oral contraceptives. However, the higher risk associated with third-generation progestins has been questioned on the basis of the possible confounding effects of age and other prescribing biases.178

Estrogenic compounds also increase the risk of VTE when they are used for lactation suppression,166 in treatment of carcinoma of the prostate, and as postmenopausal replacement therapy.179 Although estrogen doses used for postmenopausal replacement therapy are approximately one sixth those in oral contraceptives, some data support an increased thromboembolic risk at these doses as well. Several studies show a twofold to fourfold higher risk of VTE among women taking hormone replacement therapy.179183 This increased risk is greatest during the first year of treatment.180,183 However, given the relative infrequency of thromboembolism, this risk represents only one or two additional cases of thromboembolism per year in every 10,000 women in this age group.

Estrogen in pharmacologic doses is associated with alterations in the coagulation system that may contribute to this thrombotic tendency. Such alterations include decreases in PAI-1184 and increases in blood viscosity, fibrinogen, plasma levels of factors VII and X, and platelet adhesion and aggregation.185,186 An associated prothrombotic state is implied by rises in markers of activated coagulation occurring in conjunction with elevations of circulating factor VIIa and decreases in antithrombin and protein S inhibitor activity.185,187 The extent to which antithrombin and protein S are depressed is significantly less with lower estrogen preparations.188

Blood Group

There also appears to be a relationship between VTE risk and the ABO blood groups, with a higher prevalence of blood type A and correspondingly lower prevalence of blood type O groups.189,190 In reviewing the literature, Mourant et al191 found the relative incidence of type A to be 1.41 times higher among patients with thromboembolism than among controls. The effect of blood type was greater in young women who were taking oral contraceptives or were pregnant; the relative incidence of type A among patients with thromboembolism was 3.12 in those taking oral contraceptives and 1.85 in those who were pregnant. A relationship between soluble endothelial cell markers and ABO blood group is known to exist, with significantly lower levels of von Willebrand factor among those with type O blood.192

Geography and Ethnicity

There are also geographic differences in the frequency of VTE; the incidence of postoperative DVT in Europe has been noted to be nearly twice that in North America.87,92 Higher rates of thromboembolism have also been noted in the central United States compared with either coast.193 Autopsy series suggest that although the prevalence of thromboembolism is identical among American black and white patients, it is significantly higher than in a matched Ugandan black population.194 A similar autopsy series noted the prevalence of thromboembolism to be 40.6% in Boston and 13.9% in Kyushu, Japan.195

Unfortunately, regional variations in underlying medical and surgical conditions as well as in prophylactic measures and diagnostic methods may confound any apparent differences in the incidence of thromboembolism among different ethnic and geographic groups. Nevertheless, it is certainly conceivable that differences between ethnic groups might arise from either genetic or environmental factors. Such differences seem likely on the basis of recognized geographic differences in the spectrum of mutations leading to congenital anticoagulant deficiencies.196 Such theoretical concerns are also supported by geographic variability in the incidence of the factor V Leiden mutation. The factor V Leiden allele has a prevalence of 4.4% in Europeans, corresponding to a carrier rate of 8.8%, but the allele has not been identified in Southeast Asian or African populations.121

Central Venous Catheters

The use of central venous cannulation for hemodynamic monitoring, infusion catheters, and pacemakers has been associated with a rising frequency of DVT. This is particularly true for upper extremity thrombi, as many as 65% of which are related to central venous cannulation.197 Although the incidence of symptomatic thrombosis may be low, studies employing objective surveillance have reported thrombosis to occur with a mean incidence of 28% after subclavian cannulation.198 This risk also extends to femoral venous catheters; ipsilateral thrombosis develops in 12% of patients undergoing placement of large-bore catheters for trauma resuscitation.199

Experimental studies in rats suggest that the catheter itself, associated vascular injury, and stasis contribute to the development of early perigraft thrombus in virtually all cases.200 Catheter material appears to be an important determinant of thromboembolism201; polytetrafluoroethylene (Teflon) and heparin-bonded catheters are associated with thrombosis in less than 10% of cases.202 Other determinants of thrombosis are catheter diameter, number of venipuncture attempts, duration of catheter placement, and composition of the infusate.198 Heit et al203 found that current or recent placement of a central venous catheter and previous placement of a transvenous pacemaker were strongly associated with upper extremity DVT, with more than a fivefold higher risk even after adjustment for comorbid conditions.203

Inflammatory Bowel Disease

Clinical series have reported VTE to complicate inflammatory bowel disease in 1.2% to 7.1% of cases.204,205 Crohn’s disease has incidence rates of 31.4 per 10,000 person-years and 10.3 per 10,000 person-years for DVT and PE, respectively. Ulcerative colitis also has a high incidence rate of 30.0 per 10,000 and 19.8 per 10,000 person-years for DVT and PE, respectively. Such thromboses frequently occur among young patients, are more common with active disease, and may affect unusual sites, such as the cerebral veins.204,205 Greater extent of colonic disease in ulcerative colitis portends a higher risk of VTE. Most cases are not associated with inherited inhibitor deficiencies, antiphospholipid antibodies, or factor V Leiden mutation. The significance of thrombocytosis and elevations of factor V, factor VIII, and fibrinogen during active episodes of inflammatory bowel disease is unclear because all are acute phase reactants.204,205

Other coagulation abnormalities, including depressed antithrombin values, elevated PAI-1 values, and presence of anticardiolipin antibodies, have been inconsistently reported.206

Buy Membership for Surgery Category to continue reading. Learn more here