Introduction

Patients with cancer are highly susceptible to venous thromboembolism (VTE), as well as arterial thromboembolic events, complications that contribute significantly to the morbidity and mortality of the disease. There is an abundance of epidemiologic evidence to support the close-knit relationship between thrombosis and cancer. The risk of an episode of idiopathic VTE within 6 to 12 months of a new cancer diagnosis approaches 10%. Many case-control studies comparing VTE risk in cancer patients with the risk in noncancer patients consistently show that cancer patients are sixfold to sevenfold more likely to develop VTE, and among some patient groups, the risk of cancer-associated thrombosis is higher than 15%. VTE events that occur in cancer patients include deep vein thrombosis (DVT), pulmonary embolism (PE), and visceral vein thrombosis. Arterial events include myocardial infarction (MI) and stroke.

VTE has a significant impact on cancer patients. Thrombosis is the second leading cause of mortality in cancer patients after cancer itself. Cancer patients have higher rates of VTE recurrence and bleeding complications than noncancer patients, and VTE can be associated with significant symptoms.

There are many known risk factors for cancer-associated thrombosis, but our understanding of the mechanisms responsible for the hypercoagulable state characteristic of cancer is limited. Considerable overlap exists between processes responsible for cancer growth and metastasis and thrombotic complications. These include alterations in major regulatory pathways that mediate blood coagulation, platelet–vessel wall interaction, fibrinolysis, inflammatory cytokine production, and angiogenesis.

Clinical studies looking at prevention and treatment of cancer-associated thrombosis have grown, and in recent years several studies have introduced the concept of using biomarkers to estimate risk of cancer-associated thrombosis. This review focuses on epidemiology, risk prediction, and pathogenesis of cancer-associated thrombosis, and describes optimal evidence-based approaches to the prevention and treatment of VTE in this setting.

Epidemiology

Idiopathic Venous Thromboembolism and Occult Cancer

The close relationship between tumor growth and activation of blood coagulation has been known since the days of Professor Armand Trousseau, who was one of the first to describe the clinical association between idiopathic VTE and occult malignancy. Migratory thrombophlebitis as a presenting manifestation of cancer, which has come to be known as Trousseau syndrome , is rather uncommon. However, we have become increasingly aware that unprovoked VTE can be an ominous sign of underlying occult malignancy. A systematic literature review examining 34 studies that encompassed 9516 patients with VTE reported that newly diagnosed cancer was discovered in 6.3% of patients in the first 12 months after VTE diagnosis. The period prevalence of newly diagnosed cancer was 10% in patients with unprovoked VTE and 2% in patients with provoked VTE, and of those with unprovoked VTE, 6% were diagnosed with cancer at the same time as VTE.

It remains unclear whether screening for occult malignancy at the time of unprovoked VTE presentation is beneficial. A large cancer registry study encompassing over 500,000 cancer patients reported that unprovoked VTE occurred 1.3 times more than expected in the year preceding the diagnosis of cancer. However, this increased incidence was significant only for patients with advanced incurable cancer (standardized incidence ratio, 2.3) and not for patients with localized cancer (standardized incidence ratio, 1.07). This suggests that screening all patients with unprovoked VTE will be unlikely to detect early-stage cancers for which intervention could significantly impact survival. The systematic review included 15 observational cohort studies representing 4378 patients with VTE who subsequently underwent cancer screening. Most of these studies were observational cohort studies. The Screening for Occult Malignancy in Patients with Idiopathic Venous Thromboembolism (SOME) study was a large multicenter well-designed clinical trial looking at whether the addition of screening computed tomography (CT) imaging would increase early detection of occult malignancy. More than 800 patients were randomized to a limited cancer screening approach (history, exam, complete blood counts, chemistries, routine age appropriate screening for breast, cervical, and prostate cancer) or the same plus a comprehensive CT scan of the abdomen and pelvis. In this randomized trial, extensive cancer screening did not increase the detection of occult cancer, change the mean time to detection of occult cancer, or improve cancer-related mortality. The authors concluded that the addition of CT imaging did not provide a clinically significant benefit to cancer detection.

Given the uncertain overall benefit of performing extensive cancer screening in all patients with unprovoked VTE and the potential costs and risks of such a strategy, including radiation exposure and patient anxiety, extensive screening in all patients with idiopathic VTE is not recommended. For now, a complete history, physical examination, chest radiography, and age-appropriate cancer screening are an appropriate starting point, with more focused evaluation depending on the initial findings.

Venous Thromboembolism as a Complication of Cancer

In large population-based case-control studies, patients with a known diagnosis of cancer are sixfold to sevenfold more likely to develop thrombosis than are noncancer patients, and cancer accounts for almost one-fifth of attributable thrombosis risk. However, the reported rate of thrombosis in cancer patients varies considerably. Part of this variability stems from the multitude of risk factors for cancer-associated thrombosis, such as cancer type and stage, which are further discussed later. The considerable variability in study design, duration of follow-up, time period of study, and method of VTE detection and reporting also partly accounts for the wide range in reported rates. In general, VTE rates are higher in more recent studies, possibly due to greater awareness, use of different imaging modalities, or change in cancer treatments. Studies that include patients under active cancer treatment report higher rates of VTE (4% to 12.6%) than population-based database studies (0.6% to 2.1%) in which a higher proportion of patients have a remote diagnosis of cancer. Reported VTE rate is also largely dependent on the mechanism of event detection and recording. For example, in one prospective randomized study of patients with advanced colorectal cancer, VTE was initially reported as a toxicity during treatment in only 2 of 266 patients (0.8%); a subsequent retrospective review of data for the same population identified VTE in an additional 25 patients, for an actual VTE rate of 10.2%. Autopsy studies have reported VTE rates as high as 30% in cancer patients.

There are many known risk factors for cancer-associated thrombosis, including cancer-specific, patient-specific, and treatment-specific factors. The type of cancer is a particularly important factor. The rate of cancer-associated thrombosis varies from less than 1% to 12% in large studies that pool data for patients with many different types of cancers ( Table 23.1 ). In these pooled analyses, certain cancer types such as pancreatic, gastric, brain, ovarian, and lung cancer consistently have the highest VTE rates, whereas prostate and breast cancer tend to have lower rates. In addition, advanced stage and adenocarcinoma histology are associated with increased risk.

TABLE 23.1
Rates of Cancer-Associated Thrombosis in Large Pooled Studies
Study Cancer Stage Type of Study Years of Study Population Number of Patients Median Follow-Up Incidence (%)
Levitan, et al. NA Retrospective cohort 1984–1990 Hospitalized Medicare 1,211,944 NA 0.6
Sallah, et al. I–IV Retrospective cohort 1993–2000 Hospitalized with solid tumors 1,041 26 months 7.8
Stein, et al. NA Retrospective population-based cohort 1979–1999 Hospitalized 40,787,000 NA 2.0
Agnelli, et al. I–IV Prospective cohort 1999–2002 Oncologic surgery 2,373 35 days 2.1
Khorana, et al. NA Retrospective population-based cohort 1995–2002 Hospitalized neutropenic 66,106 8 days 6.4
Chew, et al. I–IV Retrospective record linkage cohort 1993–1995 Hospitalized 235,149 24 months 1.6
Blom, et al. I–IV Retrospective record linkage cohort 1986–2002 Various 66,329 6 months 3.2
Khorana, et al. NA Retrospective cohort 1995–2003 Hospitalized 1,015,598 8 days 4.1
Khorana, et al. I–IV Prospective cohort 2002–2005 Ambulatory, receiving chemotherapy 4,066 73 days 2.2
Khorana, et al. I–IV Retrospective cohort 2005–2008 Ambulatory, receiving chemotherapy 17,284 12 months 12.6
Mandala, et al. I–IV Prospective randomized 2000–2010 Chemotherapy Phase 1 1,415 NA 4.0
NA, Not available.

Patient-related demographic features such as age, race, and sex also contribute to cancer-associated risk. In the hospital and postsurgical settings, older cancer patients were more likely to experience VTE. Several studies have suggested that VTE rates are higher in African Americans but lower in people of Asian background. One study has shown higher VTE risk in women with cancer than in men, but most studies show no impact of sex on VTE risk.

Medical comorbidities clearly are associated with increased risk of thrombosis in cancer patients. In a large cohort of lung cancer patients receiving chemotherapy, a high score on the Charlson comorbidity index and the presence of congestive heart failure were associated with higher VTE risk after adjustment for other known risk factors. Increased comorbidities are a risk factor in patients with ovarian, breast, and colorectal cancer as well. Immobility, which leads to venous stasis, has long been recognized as a risk factor for VTE. In cancer patients, performance status is a clinical assessment tool that is widely used to assess mobility. In a prospective study, 31% of lung cancer patients with poor performance status while undergoing chemotherapy had VTE, compared with 15% with better performance status. Obesity is also a well-documented risk factor for cancer-associated thrombosis.

Venous Thromboembolism and Cancer Treatments

Historically, patients with cancer who underwent surgery were at an approximately twofold increased risk of postoperative thrombosis compared with noncancer patients who underwent the same operations. However, in more recent years, surgery has not been found to be a major risk factor for VTE relative to other factors, which may be due to increased awareness of the issue and use of thromboprophylaxis. Rates of VTE in association with major oncologic surgical procedures have also remained relatively stable, whereas sharp increases have been observed in patients receiving chemotherapy. Despite these improvements, there is still considerable VTE risk associated with oncologic surgery. In one large prospective study enrolling more than 2000 patients, 2.3% developed VTE, with 40% of events diagnosed more than 21 days after surgery. Risk factors for postoperative VTE included age older than 60 years (odds ratio [OR], 2.6; 95% confidence interval [CI], 1.2 to 5.7), previous VTE (OR, 6.0; 95% CI, 2.1 to 16.8), advanced cancer (OR, 2.7; 95% CI, 1.4 to 5.2), anesthesia lasting longer than 2 hours (OR, 4.5; 95% CI, 1.1 to 19), and postoperative bed rest longer than 3 days (OR, 4.4; 95% CI, 2.5 to 7.8).

Use of erythropoiesis-stimulating agents (ESAs) is known to increase thrombotic risk, particularly among patients with cancer. This was first recognized in studies of women with advanced cervical and breast cancer treated with ESAs, in whom the thrombosis rates were 13% and 29%, respectively. A systematic review encompassing 37 trials and 6743 patients showed that use of ESAs significantly increased the risk of thrombotic complications in cancer patients (relative risk [RR], 1.7). Interestingly, blood transfusions also appear to increase thrombosis risk. In a large retrospective discharge database study in multivariate analysis, red blood cell transfusion (OR, 1.60 for venous and 1.53 for arterial) and platelet transfusion (OR, 1.20 for venous and 1.55 for arterial) were independently associated with an increased risk of venous and arterial thromboembolism.

Other hematopoietic growth factors have not been as well studied. Barbui and colleagues conducted a meta-analysis of studies in which patients with cancer were treated with hemato­poietic growth factors. Although the analysis was not conclusive, the authors suggested that an increased risk of thrombosis was associated with the use of granulocyte-macrophage colony-stimulating factor. These findings have not been corroborated in newer cohort studies.

Conventional cytotoxic chemotherapy is associated with a twofold to sixfold increased risk of VTE compared with the rate in the general population. Specific chemotherapeutic agents may be associated with higher rates of VTE. In one prospective study, platinum-based regimens were significantly associated with VTE ( P = .03). Even within this class of agents, rates are higher for cisplatin than for oxaliplatin. Another study of patients receiving cisplatin-based chemotherapy reported a VTE rate of 18%.

Antiangiogenic agents have been noted to carry a particularly high risk of cancer-associated thrombosis. Bevacizumab is a humanized monoclonal antibody against vascular endothelial growth factor (VEGF) that has been shown to be effective in the treatment of several cancers, including colorectal and lung cancer. Several large clinical trials and meta-analyses have demonstrated that bevacizumab treatment is associated with approximately a twofold increased risk of arterial thromboembolic events. Bevacizumab use may also be associated with increased risk of VTE events, although this is controversial because results have been conflicting. An early meta-analysis including five clinical trials reported a significant increase in arterial thrombotic events but no difference in VTE. A later comprehensive meta-analysis examining 15 trials encompassing almost 8000 patients reported that bevacizumab therapy was associated with an increased risk of VTE (RR, 1.3), but this analysis did not adjust for the increased duration of follow-up in bevacizumab-treated patients. Pooled analysis of individual prospectively collected patient data from many clinical trials reported no increase in risk of VTE in treatment using bevacizumab. Oral VEGF tyrosine kinase inhibitors such as sunitinib and sorafenib have been shown to increase the risk of arterial thrombotic complications (RR, 3.0; 95% CI, 1.25 to 7.37; P = .015) but not the risk of VTE.

Immunomodulatory drugs such as thalidomide and lenalidomide have been widely used in the treatment of patients with multiple myeloma. These medications, especially when combined with other therapies such as high-dose corticosteroids, are associated with very high rates of thrombotic complications. Reported rates of VTE for thalidomide combination treatment are in the range of 12% to 28%, and VTE rates are also high for lenalidomide-based therapy, ranging from 5% to 75%.

Finally, patients with cancer who have indwelling central venous catheters (CVCs) are believed to be at increased risk of thrombosis of the axillary-subclavian vein, with incidence rates of up to 35% reported in the older literature. Rates appear to be lower in newer studies. In a prospective study of more than 400 cancer patients with CVCs, 4.3% developed symptomatic CVC-related DVT. Many catheter-related thromboses may be asymptomatic, as suggested by one review which noted that the incidence of symptomatic catheter-related DVT in adult patients ranges from 0.3% to 28%, whereas the rate of catheter-related DVT assessed by venography is 27% to 66%. Risk factors associated with CVC-related DVT include more than one insertion attempt (OR, 5.5; 95% CI, 1.2 to 24.6), previous CVC insertion (OR, 3.8; 95% CI, 1.4 to 10.4), left-sided placement (OR, 3.5; 95% CI, 1.6 to 7.5), catheter tip position in the superior vena cava compared with the right atrial placement (OR, 2.7; 95% CI, 1.1 to 6.6), and use of an arm port compared with a chest port (OR, 8.1; 95% CI, 3.5 to 19.1).

Clinical Impact of Cancer-Associated Venous Thromboembolism

Patients with cancer who develop VTE have a shorter life expectancy. In a population-based study reported by Silverstein and colleagues, the presence of cancer was an independent predictor of worse survival in patients who had acute VTE. Similarly, Sørensen and colleagues showed that patients with cancer who concurrently had clinically apparent VTE experienced a worse survival rate than those with cancer who did not have VTE. Levitan and associates, after analysis of the Medicare database, reported that the mortality rate for patients with VTE and malignant disease was substantially higher than that for patients with malignant disease alone. In another study of cancer patients receiving chemotherapy in an outpatient setting, thrombosis was the second leading cause of death in cancer patients following only cancer progression.

Patients with cancer who have established VTE are more likely to develop recurrent VTE during oral anticoagulation than are patients without cancer. In a trial reported by Levine and associates, patients with proximal DVT were randomly assigned to receive either unfractionated heparin (UFH) administered by continuous intravenous infusion in the hospital or low molecular weight heparin (LMWH) given subcutaneously at home, followed by 3 months of oral anticoagulant therapy. No difference was detected in the rate of recurrent thromboembolism between treatment groups. However, the rate of recurrent VTE in 103 patients with cancer was 14%, whereas the rate was 4% in 317 patients without cancer ( P = .001). In a similarly designed study reported by the Columbus investigators that compared LMWH with UFH for treatment of proximal DVT, rates of recurrent thromboembolism were 8.6% in patients with cancer and 4.1% in patients without cancer. Other studies suggest that cancer patients are at increased risk of recurrent VTE, as well as bleeding complications, compared with patients without cancer; the annual risk of recurrent VTE and bleeding was reported to be 21% in patients with cancer and 12% in those without. The majority of patients with cancer-associated VTE are hospitalized, with the mean cost of hospitalization exceeding $20,000.

Prediction of the Risk of Cancer-Associated Thrombosis

Prediction of risk of VTE has become an important clinical and research question. Although it is unquestionably true that cancer patients are at high risk of VTE, the risk varies greatly according to various cancer-, treatment-, and patient-related factors. Approaches to risk stratification are based on the following: (1) clinical variables discussed previously, such as primary site of cancer, use of chemotherapy or specific agents or regimens (e.g., immunomodulatory drug–based regimens in myeloma), or stage of disease; (2) biomarkers; and (3) risk assessment tools.

Incidental Venous Thromboembolism in Cancer Patients

Many cancer-associated VTEs are found incidentally on radiographic studies performed for cancer staging purposes. The nomenclature and classification of these events varies in the literature, but often these VTE events are referred to as incidental or clinically unsuspected VTE. Some studies suggest that more than 50% of cancer-associated thrombotic events are found incidentally on radiographic studies performed for routine staging. A 2013 study of lung cancer patients showed that one-third of VTE events and 50% of PE events were incidentally discovered, and another study of patients with multiple tumor types found that 44% of VTEs were so discovered.

The prevalence of incidental PE on routine radiographic studies of the chest performed for cancer staging or monitoring of other disease such as pulmonary nodules is 1.5% to 3.4% per scan. The rate of incidental PE rises to 9% in hospitalized cancer patients undergoing routine scanning. A study by Cronin and colleagues evaluating routine cancer staging scans of the chest, abdomen, and pelvis found unsuspected iliofemoral DVT in 6.8%, unsuspected common iliac DVT in 1.2%, unsuspected inferior vena cava (IVC) DVT in 0.3%, and unsuspected PE in 3.3% of patients.

A retrospective study of 403 cancer patients undergoing routine multidetector thoracic CT suggests that many incidental VTE events may initially go unnoticed. In this study, all scans were reviewed independently by two radiologists and all PEs were confirmed by three thoracic radiologists. Four percent of patients (16 of 403) were found to have incidental PEs, yet only 25% of these were noted on the initial radiology report. In this series, none of the PEs was treated, and all 12 were found to have resolved without benefit of anticoagulation on subsequent scans performed a median of 4.7 months after the index scan.

Although unsuspected, many incidentally discovered VTEs are not truly asymptomatic, and these events may also have a negative impact on survival. An important case-control study in which cancer patients with unsuspected PE were compared with matched cancer patients without PE showed that patients with unsuspected PE were significantly more likely to report fatigue (54% vs. 20%) and shortness of breath (22% vs. 8%) despite no difference in the prevalence of anemia or lung metastases among the two groups. A 2014 analysis showed similar rates of VTE recurrence, bleeding, and mortality in cancer patients with incidentally discovered VTE compared with cancer patients with suspected VTE. The clinical importance of incidental VTE is also supported by a 2011 multivariate analysis of data comparing cancer patients with incidental VTE to matched cancer patients without VTE. In that study, patients with unsuspected PE were found to have significantly abbreviated survival (hazard ratio [HR], 1.51; 95% CI, 1.01 to 2.27, P = .048). That study also showed that clot burden may be important, because survival in patients with only isolated subsegmental unsuspected PE was not significantly different from that of matched controls (HR, 1.04; 95% CI, 0.44 to 2.39; P = .92). A study involving lung cancer patients also showed that both incidental VTE (HR, 2.4) and suspected VTE (HR, 2.7) were associated with shortened survival.

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here