Cancer-associated hypercoagulable state: Causes and mechanisms

INTRODUCTION — The pathogenesis of thrombosis in malignancy involves an underlying prothrombotic state resulting from cancer-related procoagulant mechanisms, which is further aggravated by clinical risk factors that occur more commonly in cancer patients, such as hospitalization, surgery, catheter insertion, and cancer-directed therapy.

The causes, mechanisms, and clinical risk factors of cancer-associated thrombosis will be reviewed here.

Prevention and treatment of hypercoagulable disorders associated with malignancy are discussed separately.

●Prevention

•Multiple myeloma – (See "Multiple myeloma: Prevention of venous thromboembolism".)

•Pancreatic cancer – (See "Supportive care for locally advanced or metastatic exocrine pancreatic cancer", section on 'Venous thromboembolism'.)

•Other tumors – (See "Risk and prevention of venous thromboembolism in adults with cancer", section on 'Primary prevention'.)

●Treatment

•Catheter-related thrombosis – (See "Catheter-related upper extremity venous thrombosis in adults", section on 'Treatment'.)

•Deep vein thrombosis and pulmonary embolism – (See "Anticoagulation therapy for venous thromboembolism (lower extremity venous thrombosis and pulmonary embolism) in adult patients with malignancy".)

ASSOCIATION BETWEEN CANCER AND THROMBOSIS — Observations of an association between cancer and an increased risk of thrombosis were made by Armond Trousseau in the mid-1800s. Over 150 years later, it is now well established that thrombosis occurs commonly in cancer patients. Trousseau died of gastric cancer in 1867.

●The cancer-associated hypercoagulable state may present clinically as venous thromboembolism (VTE), arterial thromboembolism, microcirculatory thrombosis, or superficial thrombophlebitis. The risks of venous and arterial thromboembolism are increased several-fold, as discussed separately. (See "Risk and prevention of venous thromboembolism in adults with cancer", section on 'Incidence and risk factors'.)

●Certain types of cancers, especially pancreatic, gastric, brain, and ovarian cancers, are associated with a particularly high incidence of thrombosis. Thrombotic complications typically present in close temporal relationship with the diagnosis of cancer but may also precede the diagnosis by months. (See 'Cancer-related factors' below.)

●Other factors, including cancer therapy, patient age, and comorbidities influence the thrombotic risk in patients with cancer. (See 'Patient-related factors' below and 'Therapy-related factors' below.)

BIOMARKERS — A number of potential biomarkers have been postulated to be useful for identifying patients with cancer at high risk for venous thromboembolism (VTE) and other types of thrombosis, mostly based on mechanistic studies. (See 'Molecular mechanisms' below.)

However, the only biomarkers that have been extensively validated and widely applied to clinical care are the complete blood count (CBC) parameters as elements of the Khorana risk prediction model for VTE in cancer (table 1). (See "Risk and prevention of venous thromboembolism in adults with cancer".)

●High platelet count – ≥350,000/microL before chemotherapy

●Low hemoglobin – <10 g/dL or use of an erythropoiesis-stimulating agent (ESA)

●High white blood cell (WBC) count – >11,000/microL before chemotherapy

D-dimer is a fibrin degradation production released upon cleavage of cross-linked fibrin. Elevations in D-dimer can be indicative of increased thrombin generation due to thrombosis or an underlying prothrombotic state. D-dimer has not been validated as a biomarker for cancer-associated VTE in the same way as the Khorana score, although it appears to correlate with VTE in patients with cancer. In non-cancer populations, D-dimer has high sensitivity and low specificity for VTE. (See "Clinical presentation and diagnosis of the nonpregnant adult with suspected deep vein thrombosis of the lower extremity", section on 'D-dimer'.)

D-dimer values are often higher in cancer cohorts than those without cancer [1]. More extreme elevations (above the 75^th percentile) have been linked to an increased risk of VTE [2]. D-dimer values can fluctuate over time, and a rising trajectory may improve predictive accuracy [3]. Due to issues with low positive predictive value, unresolved D-dimer cutoffs in cancer, and fluctuating values, D-dimer values are not routinely used in clinical practice to predict thrombosis in individuals with cancer.

There is keen interest in establishing better and more specific mechanistic biomarkers, such as tissue factor, soluble P-selectin, and neutrophil extracellular traps, but none of these has been validated sufficiently for clinical care. (See 'Molecular mechanisms' below.)

MOLECULAR MECHANISMS — The hypothesized pathogenesis of venous thrombosis is referred to as Virchow's triad; it consists of:

●Hypercoagulability

●Decrease in venous flow (or stasis)

●Vessel wall damage (or endothelial injury)

In individuals with cancer, these can be due to the malignancy itself, cancer treatment, and factors not directly related to the cancer.

Procoagulant proteins — Cancer cells may express or secrete procoagulant proteins that directly activate the coagulation cascade in the circulation.

●Tissue factor – Tissue factor (TF) is a transmembrane protein that plays a critical role in the initiation of coagulation and hemostasis. (See "Overview of hemostasis", section on 'Tissue factor'.)

In physiologic conditions, TF is expressed on perivascular and epithelial cells. Expressed TF binds to activated factor VII (FVIIa), and the TF-FVIIa complex activates factor IX and factor X, triggering the downstream coagulation cascade. (See "Overview of hemostasis", section on 'Activation of factors VII, X, and IX'.)

Certain types of tumor cells demonstrate high levels of TF expression. Such cells are capable of releasing TF into the circulation in the form of TF-positive extracellular vesicles, contributing to the development of cancer-associated thrombosis [4,5]. Studies using immunohistochemistry have observed high levels of TF expression predominantly in pancreatic cancer cells [6-8]. Expression has also been seen in other cancer types including cervical, brain, and ovarian cancer [9-11].

TF expression may correlate with tumor histological grade and angiogenesis [7,8,12]. In a retrospective analysis of 41 patients with pancreatic cancer, VTE was more common among patients with high TF-expressing carcinomas than those with low TF-expressing carcinomas (26 versus 5 percent) [8].

Certain oncogenic gene mutations appear to correlate with upregulation of TF expression:

•KRAS and TP53 mutation in colorectal cancer [13]

•PTEN, TP53, KRAS, and ALK mutation in non-small cell lung cancer [14,15]

•EGFR overexpression and IDH1 wild-type status in glioma [16-18]

In addition to intratumoral expression, TF can be carried through the circulation by cancer cell-derived extracellular vesicles [4,19]. These TF-positive extracellular vesicles (TF+ EVs) can be measured in plasma using enzyme-linked immunosorbent assays (ELISA), flow cytometry, or activity-based assays [20].

Although the definitive role of TF+ EVs in clinical VTE remains undetermined in other cancers, several cohort studies have consistently observed an association between plasma TF levels and VTE in pancreatic cancer [4,21-24]. In one study in pancreatic cancer, TF+ EVs were associated with increased mortality but not VTE [25]. The concentration of TF+ EVs can be nearly eliminated after curative pancreatectomy [4].

●Podoplanin – Podoplanin (PDPN) is a transmembrane protein that activates platelets via binding to platelet C-type lectin-like receptor 2 (CLEC-2) [26,27]. PDPN is expressed on lymphatic endothelial cells and kidney podocytes. It is upregulated in various tumor cells including squamous cell carcinoma, germinal tumors, mesothelioma, and brain tumors [28].

Some studies suggest that PDPN plays a role in cancer-associated thrombosis. In an immunohistochemistry study of primary brain tumors, high PDPN expression was associated with an increased risk of VTE [29]. In a mouse model of ovarian cancer, PDPN-positive EVs were associated with platelet aggregation and VTE [30].

●PAI-1 – Plasminogen activator inhibitor (PAI-1) is the principal inhibitor of plasminogen activators and plasma fibrinolytic activity. (See "Overview of hemostasis", section on 'Plasminogen activator inhibitors and alpha-2-antiplasmin'.)

PAI-1 levels are elevated in pancreatic cancers [31]. In a study involving 139 pancreatic cancer patients, increased levels of active PAI-1 were independently associated with VTE, with a 40 percent increase in VTE risk per doubling of PAI-1 level [32]. The same study showed that mice bearing pancreatic tumors with PAI-1 overexpression had impaired clot resolution eight days after ligation of the inferior vena cava. These data suggest that elevated PAI-1 and decreased fibrinolysis contribute to the pathogenesis of cancer-associated thrombosis.

Conversely, increased expression of the plasminogen receptor annexin A2 on promyelocytic leukemia cells contributed to hyperfibrinolysis and an underlying coagulopathy [33,34].

●PDI – Protein disulfide isomerase (PDI) is a thiol isomerase localized in the endoplasmic reticulum that has an essential role in protein folding and disulfide bond formation during protein synthesis. PDI is released from activated platelets or endothelial cells. (See "The endothelium: A primer", section on 'Procoagulant properties' and "Platelet biology and mechanism of anti-platelet drugs", section on 'Secretion (granule exocytosis)'.)

Extracellular PDI may contribute to hypercoagulability in cancer through interactions with different coagulation factors [35,36]. In a study of 65 patients with myeloproliferative neoplasms (MPNs) and 27 healthy controls, elevated plasma PDI levels were detected primarily in JAK2-mutated MPNs and were predictive of VTE and the risk of arterial thromboembolism [37]. (See "Overview of the myeloproliferative neoplasms", section on 'JAK2 mutations' and "Clinical manifestations and diagnosis of polycythemia vera", section on 'Thrombosis and hemorrhage'.)

Quercetin flavonoids, found in fruits and vegetables, have been identified as PDI inhibitors. In a phase 2 trial in patients with advanced colorectal, pancreatic, or lung cancer, inhibition of PDI with isoquercetin significantly reduced the levels of plasma D-dimer, soluble P-selectin, and platelet-dependent thrombin generation [38]. (See "Investigational anticoagulants", section on 'Protein disulfide isomerase inhibitors'.)

Endothelial cells, WBCs, and platelets — Tumor cells may express proteins that interact with non-cancer cells and promote thrombosis. Elevated white blood cell (WBC) count and platelet count are two of the biomarkers validated for clinical use as part of the Khorana score. (See 'Biomarkers' above.)

●Endothelial cells – Inflammatory cytokines released from cancer cells, such as tumor necrosis factor (TNF-) alpha, interleukin (IL-) 1beta, IL-6, IL-8, and vascular endothelial growth factor (VEGF), can induce a procoagulant phenotype in endothelial cells. The phenotype includes expression of TF, release of von Willebrand factor (VWF), and release of plasminogen activator inhibitor 1 (PAI-1) [39]. (See 'Procoagulant proteins' above.)

●WBCs

•Leukocytosis – Elevated WBC counts are frequently observed in patients with cancer and are associated with VTE in this population [40-42]. In the derivation cohort of the Khorana score, pre-chemotherapy WBC count >11,000/microL was associated with doubling in the risk of VTE (odds ratio [OR] 2.2, 95% CI 1.2-4) [43]. Analysis of a cohort from the Vienna Cancer and Thrombosis Study (CATS) showed a 7 percent increase in the risk of VTE per 1000/microL increase in the WBC count [42].

•Neutrophil extracellular traps (NETs) – NETs are extracellular decondensed chromatin fibers decorated with granular proteins and histones. NETs are released from activated or dying neutrophils, mediated by the enzyme peptidylarginine deiminase 4 (PADI4) [44].

NET formation (NETosis) promotes thrombosis by providing a scaffold for platelet and fibrin deposition, activating factor XIIa, binding to VWF, and inactivating tissue factor pathway inhibitor (TFPI) [44,45].

Biomarkers associated with NETs include citrullinated histone H3 (H3Cit), myeloperoxidase, nucleosomes, neutrophil elastase, and cell-free DNA (cfDNA) [46]. In one cohort, elevated levels of H3Cit were associated with an increased risk of VTE in cancer patients for up to two years, specifically in individuals with cancer of the pancreas, lung, or breast [47]. Elevated levels of cfDNA and nucleosomes were also associated with increased risk of VTE in the first three to six months following the diagnosis of cancer or disease progression. The magnitude of risk was comparable with previously identified biomarkers for VTE in patients with cancer, such as D-dimer, soluble P-selectin, and factor VIII levels [47]. Questions have been raised regarding the sensitivity and specificity of different NET assays [48].

•Monocytes – Activated monocytes can express TF on their surface and can be a major source of procoagulant TF in the circulation. (See 'Procoagulant proteins' above.)

Increased monocyte count has been shown as a risk factor for VTE in cancer patients [49,50]. In one study in ambulatory cancer patients, an absolute monocyte count >1200/microL was associated with an increased risk of VTE [49]. In another study of hospitalized cancer patients, a monocyte count >500/microL was associated with a 5-fold increased risk of VTE [50].

●Platelets – Thrombocytosis, identified prior to or at the time of cancer diagnosis, is associated with VTE in cancer patients [40,51,52]. Different cutoffs for defining thrombocytosis were used in these studies (platelet count of >443,000/microL, ≥350,000/microL, or ≥295,000/microL); the Khorana score uses ≥350,000/microL before chemotherapy. (See 'Biomarkers' above.)

In another study, an inverse relationship was found in patients with high-grade glioma, whereby low platelet count was associated with an increased risk of VTE (hazard ratio 0.73 for each 50,000/microL increase in platelet count, 95% CI 0.53-0.95) [53].

Activated platelets express P-selectin, and increased levels of circulating soluble P-selectin have been shown to be associated with VTE in cancer populations [53-55].

In a general cancer population study, other biomarkers of platelet activation such as soluble CD40 ligand, thrombospondin-1, and platelet factor-4 (PF4) were not associated with increased risk of VTE [55]. However, in a subgroup with pancreatic cancer, high PF4 levels were associated with a 2.7-fold increased risk of VTE.

CONTRIBUTING FACTORS — The risks of thrombosis in individual cancer patients are influenced by numerous factors. Some are unique to cancer, and others may overlap with non-cancer populations (table 2).

Cancer-related factors

●Type – Cancer type is one of the most important risk factors for cancer-associated thrombosis. The highest risks of VTE are observed in pancreatic, stomach, brain, colorectal, lung, and ovarian cancers [56-58].

In the Khorana score, stomach and pancreatic cancers are considered very high risk (odds ratio [OR] 4.3, 95% CI 1.2-15.6) and are each assigned two points [43]. (See "Risk and prevention of venous thromboembolism in adults with cancer", section on 'VTE risk assessment/Khorana score'.)

Breast and prostate cancers, which are more common, have relatively lower risks of VTE [57].

●Stage – VTE risk increases with more advanced cancer stage. In one study, regional disease had a 3.7-fold higher risk of VTE than local disease [59]. Metastatic disease at the time of diagnosis is one of the strongest risk factors for VTE and is associated with 1.4- to 21.5-fold higher risk of VTE depending on cancer type [60].

●Grade – Histological grade of a tumor influences the risk of VTE. In one study, patients with high-grade tumors had twice the VTE risk of those with low-grade tumors [61].

●Timing relative to diagnosis – Risks of cancer-associated thrombosis are highest during the period immediately following the diagnosis of cancer, probably due to the active disease status, the intensity of treatments received, and associated hospitalization and procedures [57,62,63]. The risk of VTE following diagnosis gradually decreases over time [57].

•VTE – One study found that the incidence of VTE started to increase over that expected four months prior to the diagnosis of certain cancers (acute myeloid leukemia, non-Hodgkin lymphoma, and pancreatic, ovarian, stomach, renal cell, and lung cancer) [64].

•Arterial thromboembolism – One study found that the risk of arterial thrombosis progressively increased starting 150 days before the diagnosis of cancer and peaked during the 30 days immediately before diagnosis [63]. At six months after diagnosis, the risk of arterial thromboembolism doubled compared with controls; the excess risk evens out with controls at approximately one year after diagnosis [62].

●Gene variants – In some cancer types, specific pathogenic variants in certain genes influence the risks of VTE.

•Increased risks of VTE are associated with JAK2 V617F mutation in myeloproliferative neoplasms (MPNs) [65,66], ALK mutation or ROS1 rearrangement in lung cancer [67-69], and wild-type IDH1 in brain tumors [70].

•In a genomic analysis of solid tumors, tumor-specific mutations in KRAS, STK11, KEAP1, CTNNB1, CDKN2B, and MET were associated with increased risks of VTE, whereas mutations in SETD2 were associated with a decreased risk of VTE [71].

Patient-related factors

●Age – As in the general population, the risks of VTE generally increase with age in cancer patients, although the effect is quite modest (for age ≥65 years, the odds ratio in one study was 1.08, 95% CI 1.05-1.1) or is even not captured in some studies [57,60,72]. In the more highly thrombotic cancer types (pancreas, stomach) as well as lung cancer and mesothelioma, the effect is reversed, so that younger age is associated with higher risks of VTE [57,60].

●Hereditary thrombophilia – Factor V Leiden and other prothrombotic genetic variants, including prothrombin G20210A and variants in the ABO, FGG, and F11 genes have been shown to confer additional risk of VTE in patients with cancer [73-75].

TiC-Onco is a risk assessment model that incorporates a genetic risk score using four genetic variants with other clinical variables [76]. The score has comparable performance to the Khorana score, although further validation is needed.

●Other VTE risk factors – Patients with cancer may also have other comorbidities or conditions that are known to increase VTE risk, such as prior history of VTE, high body mass index (BMI), family history of VTE, or acute illness [77]. Of these, prior history of VTE is the most important risk factor, associated with an eight-fold increased risk of VTE [56].

●DIC – We do not consider disseminated intravascular coagulation (DIC), especially when it is acute, to be a mechanism of VTE similar to those described above.

However, chronic DIC in patients with cancer carries a high risk of thrombosis. (See "Evaluation and management of disseminated intravascular coagulation (DIC) in adults", section on 'Chronic DIC'.)

●COVID-19 – Relatively high incidences of VTE have been observed in hospitalized patients with coronavirus disease 2019 (COVID-19), and patients with cancer may be at higher risk of developing VTE during COVID-19 infection compared with non-cancer patients [78-80]. Among cancer patients hospitalized with COVID-19, VTE risk was increased in those with recent cancer therapy, VTE history, and/or direct intensive care unit (ICU) admission [81].

The CoVID-TE risk assessment model was proposed as a risk-stratification tool for VTE in cancer populations [81]. The model stratifies hospitalized patients with cancer and COVID-19 into low-risk and high-risk cohorts, with a corresponding VTE risk of 4.1 versus 11.3 percent.

Therapy-related factors — In a 2021 population-based study, receiving cancer treatment during the first four months after a cancer diagnosis was identified as a risk factor for VTE [56]. Cancer treatments that increased VTE risk included targeted therapy (hazard ratio [HR] 3.85, 95% CI 3.43-4.32), chemotherapy (HR 3.35, 95% CI 3.06-3.66), surgery (HR 2.20, 95% CI 2.02-2.39), and radiotherapy (HR 2.16, 95% CI 1.94-2.39) [56]. Risk was not significantly increased with hormonal therapy in this study.

Chemotherapy — Chemotherapeutic drugs have been identified as an important independent risk factor for VTE in cancer patients, increasing the risk of VTE three-fold compared with no treatment and 24-fold compared with matched controls without cancer [56]. Analysis of the use of chemotherapy showed that the proportion of cancer patients receiving chemotherapy during the first four months after cancer diagnosis has doubled from 17 percent in 1997 to 33 percent in 2017 [56].

The thrombotic risk varies with different chemotherapeutic agents.

●Cisplatin – Cisplatin-based chemotherapy has been the most notable in terms of its thrombotic risk. A systematic review and meta-analysis of 8216 patients from 38 randomized controlled trials showed a significantly increased risk of VTE in patients treated with cisplatin-based chemotherapy compared with non-cisplatin-based regimens (relative risk [RR] 1.67, 95% CI 1.25-2.23) [82]. Regimens in the control arm also included oxaliplatin- and carboplatin-based therapy.

Oxaliplatin does not appear to confer this degree of thromboembolic risk; in a randomized trial, thromboembolism occurred in 13 to 17 percent of individuals randomly assigned to receive a cisplatin-containing regimen versus 8 percent of individuals assigned to receive an oxaliplatin-containing regimen [83].

The risk of arterial thromboembolism was not found to be increased with cisplatin [84].

●Asparaginase – Asparaginase depletes the amino acid asparagine, required for synthesis of asparagine-containing proteins including antithrombin.

Prevention and management of VTE in patients receiving asparaginase are discussed separately. (See "Antithrombin deficiency", section on 'Patients receiving asparaginase'.)

Hormonal therapies

●Tamoxifen – Tamoxifen is a selective estrogen receptor modulator; it is associated with an increased risk of VTE, especially in the first years of treatment [85-89]. A prospective cohort study reported that tamoxifen initiation was associated with increased thrombin generation and increased protein C resistance from baseline as evidenced by thrombin generation assay [90]. Fibrinogen, antithrombin, protein C, and TFPI were also decreased post-treatment, effects not found with aromatase inhibitors.

Data from prevention trials in individuals without cancer also indicate that tamoxifen increases VTE risk. Examples include the NSABP P-1 Breast Cancer Prevention Trial (BCPT), which included over 13,000 patients randomly assigned to tamoxifen or placebo and showed an increased risk for VTE (risk ratio for pulmonary embolism 3.0, 95% CI 1.1-11.2, and for deep vein thrombosis 1.6, 95% CI 0.9-2.9) [91]. The International Breast Cancer Intervention Study (IBIS-1), which randomly assigned over 7000 patients to tamoxifen or placebo, found an increased risk of VTE with tamoxifen (OR 2.1, 95% CI 1.1-4.1) [92]. Risk was further increased in patients who had surgery, immobilization, or fracture in the month prior to the event, and it persisted during five years of active treatment but not following discontinuation [93].

Available data on risk of arterial thrombosis are conflicting, and a potential increased risk of stroke may be balanced by a potential decreased risk of myocardial infarction:

•In a meta-analysis from the Early Breast Cancer Trialists' Collaborative Group (EBCTCG), which compared 21,457 women randomly assigned to receive tamoxifen or control (identical chemotherapy without tamoxifen, placebo), there was a trend toward increased stroke-related mortality that did not reach the level of statistical significance (3 extra deaths per 1000 patients during the first 15 years, none of which occurred during treatment) [94]. However, there was a reciprocal trend toward decreased cardiac mortality (3 fewer deaths per 1000 patients during the first 15 years). Thus, the overall vascular mortality was similar with tamoxifen versus no tamoxifen.

•In a case-control study of 11,045 women with breast cancer, the risk of stroke was not increased by the use of tamoxifen (OR 1.0, 95% CI 0.6-1.6) [95].

•In a meta-analysis of randomized trials of tamoxifen use for breast cancer (primary or secondary prevention) that included 39,601 patients, the frequency of ischemic stroke was greater in those who received tamoxifen than controls (0.71 versus 0.39 percent; mean duration of follow-up 4.9 years) [96]. It was concluded that tamoxifen was associated with an increased stroke risk, although the absolute risk was small. This study did not evaluate myocardial ischemia.

●Raloxifene – Raloxifene is a selective estrogen receptor modulator used for prevention of osteoporosis and breast cancer. In a meta-analysis that included nearly 25,000 post-menopausal individuals without cancer, raloxifene was associated with an increased risk for VTE (OR 1.62, 95% CI 1.25-2.09) [97]. A US Preventive Services Task Force analysis comparing tamoxifen and raloxifene for breast cancer prevention found that the odds ratio for VTE was lower for raloxifene (1.56, versus 1.9 for tamoxifen) [98].

Aromatase inhibitors (anastrozole, exemestane, and letrozole) do not appear to increase the risk of VTE [99].

Targeted and immunological therapies

●Cyclin-dependent kinase inhibitors – Inhibitors of cyclin-dependent kinase 4/6, such as palbociclib, abemaciclib, and ribociclib, are associated with increased risk of thromboembolic complications. The overall one-year incidence of venous and arterial thrombosis was 10.4 percent in patients receiving these drugs [100].

●Antiangiogenic agents – Increased thrombotic risk has been observed with the anti-vascular endothelial growth factor (VEGF) monoclonal antibody bevacizumab. A systematic review and meta-analysis of 7956 patients from 15 randomized clinical trials showed an 11.9 percent incidence of VTE among patients who received bevacizumab, with an RR of 1.33 (95% CI 1.13-1.56) compared with controls [101]. (See "Cardiovascular toxicities of molecularly targeted antiangiogenic agents", section on 'Arterial and venous thromboembolism'.)

●Immunomodulatory drugs (IMiDs) and glucocorticoids – Thalidomide and lenalidomide are associated with increased risk of VTE when used in combination with chemotherapy or glucocorticoids [102-104]. (See "Multiple myeloma: Prevention of venous thromboembolism", section on 'VTE incidence and risk factors'.)

High-dose dexamethasone with lenalidomide appears to particularly increase VTE risk. (See "Multiple myeloma: Prevention of venous thromboembolism", section on 'Immunomodulatory drugs'.)

Glucocorticoids do not appear to be a major risk factor for VTE in other settings besides combination with an IMiD in patients with multiple myeloma.

●Immunotherapy and CAR-T cells – The off-target effects of T-cell activation by immune checkpoint inhibitors such as PD-L1 or CTLA4 inhibitors can cause immune-mediated vasculitis and monocyte activation, leading to heightened thrombotic risk [105]. Incidences of VTE in retrospective analyses of cancer patients who received immunotherapy ranged from 4 to 24 percent [106]. Compared with cancer patients who did not receive treatment, immune checkpoint inhibitors significantly increased the risk of VTE (adjusted HR 2.78, 95% CI 1.61-4.80) [56]. (See "Toxicities associated with immune checkpoint inhibitors", section on 'Cardiovascular toxicity'.)

Chimeric antigen receptor T cell (CAR-T) therapy may increase VTE risk. In one study, new VTE occurred in 6.7 to 11 percent within the first three months after CAR-T cell infusion [107,108]. The magnitude of increased risk compared with controls is challenging to determine. (See "Overview of gene therapy, gene editing, and gene silencing", section on 'Cancer therapy'.)

Erythropoiesis stimulating agents — Erythropoiesis stimulating agents (ESAs) are sometimes used for anemia associated with cancer. In a Cochrane systematic review of 57 trials, the risk ratio for thromboembolic complications was increased in patients receiving ESAs compared with controls (RR 1.52, 95% CI 1.34-1.74) [109]. Use of ESAs is assigned a score of 1 in the Khorana score for the prediction of cancer-associated thrombosis [43]. (See "Risk and prevention of venous thromboembolism in adults with cancer", section on 'VTE risk assessment/Khorana score' and "Role of erythropoiesis-stimulating agents in the treatment of anemia in patients with cancer", section on 'Issues related to thromboembolic risk'.)

Cancer surgery and hospitalization — (See "Risk and prevention of venous thromboembolism in adults with cancer", section on 'Inpatients (VTE prophylaxis)' and "Prevention of venous thromboembolic disease in adult nonorthopedic surgical patients".)

Indwelling catheters or ports — (See "Catheter-related upper extremity venous thrombosis in adults", section on 'Epidemiology and risk factors' and "Catheter-related upper extremity venous thrombosis in adults", section on 'Treatment'.)

DRUG-INDUCED THROMBOTIC MICROANGIOPATHY — Several cancer drugs can cause a drug-induced thrombotic microangiopathy (DITMA), as summarized in the table (table 3).

The mechanism may be immune or non-immune. (See "Drug-induced thrombotic microangiopathy (DITMA)", section on 'Drugs (immune mechanism)' and "Drug-induced thrombotic microangiopathy (DITMA)", section on 'Drugs (non-immune mechanism)'.)

SUMMARY

●Association of cancer with thrombosis – It is well established that thrombosis occurs commonly in cancer patients, including venous thromboembolism (VTE) and arterial thrombosis. The magnitude of risk and prevention strategies are discussed separately. (See 'Association between cancer and thrombosis' above and "Risk and prevention of venous thromboembolism in adults with cancer", section on 'Incidence and risk factors'.)

●Biomarkers – The only biomarkers that have been extensively validated and widely applied to clinical care are the complete blood count (CBC) parameters as elements of the Khorana risk prediction model for VTE in cancer; these include platelet count ≥350,000/microL, hemoglobin <10 g/dL (or use of an erythropoiesis-stimulating agent [ESA]), and white blood cell (WBC) count >11,000/microL (table 1). (See 'Biomarkers' above.)

●Mechanisms – Cancer cells may express or secrete procoagulant proteins that directly activate the coagulation cascade in the circulation, especially tissue factor (TF) on the cells or on circulating microparticles. Elevations in other proteins (D-dimer, podoplanin, plasminogen activator inhibitor 1 [PAI-1], protein disulfide isomerase [PDI]) and cells (platelets, WBCs, neutrophil extracellular traps [NETs]), may be surrogates for increased VTE risk or may contribute to the increased risk. We consider disseminated intravascular coagulation (DIC) to be a separate mechanism. (See 'Molecular mechanisms' above and "Evaluation and management of disseminated intravascular coagulation (DIC) in adults", section on 'Chronic DIC'.)

●Contributing factors – Factors that contribute to the hypercoagulable state in cancer include (table 2):

•Cancer type, stage, and timing – Cancer type is one of the most important risk factors for cancer-associated thrombosis. The highest risks of VTE are observed in pancreatic, stomach, brain, colorectal, lung, and ovarian cancers. Cancer stage, grade, and timing relative to diagnosis also affect VTE risk. Risk is highest from three to four months before diagnosis and may return to baseline at one year after diagnosis. (See 'Cancer-related factors' above.)

•Patient age and comorbidities – VTE risk increases with patient age and other risk factors not directly related to cancer (high body mass index [BMI], inherited thrombophilia, prior history of VTE, COVID-19). (See 'Patient-related factors' above.)

•Therapy – Chemotherapeutic agents with the highest risk of VTE include cisplatin and asparaginase. Other therapies with increased risk of VTE include tamoxifen, raloxifene, cyclin-dependent kinase inhibitors, antiangiogenic agents, immunomodulatory drugs (thalidomide and analogs), immunotherapies, and erythropoiesis stimulating agents. Surgery, hospitalization, and indwelling catheters and ports also increase VTE risk. (See 'Therapy-related factors' above.)

Risk reduction in patients receiving these is discussed separately. (See "Antithrombin deficiency", section on 'Patients receiving asparaginase' and "Cardiovascular toxicities of molecularly targeted antiangiogenic agents" and "Multiple myeloma: Prevention of venous thromboembolism" and "Risk and prevention of venous thromboembolism in adults with cancer", section on 'Primary prevention'.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Lawrence LK Leung, MD, who contributed to earlier versions of this topic review.