Thromboembolism in Gastrointestinal Cancers

Abstract

The link between thromboembolism and cancer has been recognized for over 100 years. Venous thromboembolism (VTE) is associated with considerable morbidity in patients with cancer, with emerging research also indicating a detrimental effect on survival. Investigations aimed at improving outcomes for patients with cancer have focused on the role of low molecular weight heparin in primary and secondary prevention of VTE and in improving patient survival. Important fundamental questions remain unanswered, however, and a significant line of research needs to be dedicated to investigating VTE in GI cancers. The effect of VTE on survival needs to be clarified, as does the role of anticoagulation in this patient population. Opportunities for additional research include investigating methods to identify patients at risk of developing VTE and developing new strategies and therapeutic interventions to reduce the morbidity and mortality associated with VTE. This review focuses on the current understanding of VTE related to gastrointestinal cancers and directions of interest in research specific to GI cancers and VTE.

Armand Trousseau was the first to describe the association between cancer and thrombosis when he reported cases of migratory thrombophlebitis in patients with cancer in 1867.¹ Over 125 years later, the association between cancer and thrombosis has been well established. ^2–4 The prothrombotic state generated by malignancy is multifactorial. Several molecular and cellular etiologies have been proposed, with some of the most intriguing hypotheses including expression of tissue factor by tumor cells, factor X-activating cysteine protease, and procoagulant microparticles derived from platelets, endothelial cells, and leukocytes.^5,6

Research on thromboembolism in patients with cancer has focused on improving the primary and secondary prevention of venous thromboembolism (VTE) and assessing the detrimental effect of VTE on survival and quality of life. This review summarizes the current understanding of VTE in gastrointestinal (GI) cancers and implications for future research in this area.

INCIDENCE

Gastrointestinal (GI) cancers are associated with a high incidence of thromboembolic events. In a review of over 1 million Medicare patients hospitalized with a diagnosis of malignancy, GI cancers (pancreas, stomach, liver, colon, rectum) were among the top 10 cancers with the highest rate of deep vein thrombosis (DVT) or pulmonary embolism (PE) out of 18 cancers reported (Table 1).² In addition, the risk of cancer within 1 year after idiopathic DVT or PE is elevated for pancreatic cancer, gastric cancer, esophageal cancer, colorectal cancer, and primary liver cancer.^4,7 Such retrospective data do not, however, provide sufficient information on the effects of treatment, disease stage, and use of anticoagulation for primary prophylaxis on thromboembolism risk. In a prospective series of 2,482 patients treated in clinical trials at Memorial Sloan-Kettering Cancer Center during a 2-year period, Asmis et al found a rate of VTE of 8.6% among GI cancer patients, compared with 3.3% in patients with non-GI cancers (P < .0001).⁸

Table 1.

Risk analysis using Medicare claims data.

Cancer	Rate of DVT/PE Per 10,000 patients	Rank out of 18 malignancies
Pancreas	110	3
Stomach	85	5
Colon	76	8
Liver	69	9
Rectal	62	10
Esophagus	43	15

Abbreviations: DVT = deep vein thrombosis; PE = pulmonary embolism. From Levitan et al.²

The reported incidence of VTE varies widely based on the location of the primary GI tumor, disease stage, and other factors. Although there appears to be increased awareness of the relationship between thrombosis and cancer, the reporting of VTE in prospective series and clinical trials remains inconsistent.

Advanced Disease

Although the general perception is that VTE is common in the palliative chemotherapy setting, rates have not been well defined in prospective clinical trials. The reported overall rate of VTE has been as high as 71% in patients with advanced pancreatic cancer treated with chemotherapy,⁹ although rates are infrequently reported in phase III trials.^10–14 In gastric cancer, the reported rate of VTE ranges from 5.3% to 25.5% in phase II clinical trials, with few phase III data being available.^15–19 Reported rates for VTE during palliative therapy for other GI cancers include colorectal cancer, 0.4% to 19.4%; hepatocellular carcinoma, 1.4% to 4%; and bile duct and gallbladder cancer, 2% to 16.7%.^20–24

It is likely that the true rate of VTE is in the range of 10% to 20% for patients with GI cancers in advanced stages, with the highest rates in cancers of the pancreas, stomach, and colon. In the future, with more advanced imaging technology, it is likely that reported rates of VTE will increase—albeit in the setting of increased reporting of asymptomatic VTE with unknown clinical significance. Nevertheless, it is important that diligent reporting of thromboembolic events be undertaken and maintained so that a better understanding of the natural history of this phenomenon can be gained.

Localized Disease

The reporting of VTE has been even less consistent in clinical trials in patients with localized GI cancers. Recent adjuvant and preoperative phase III trials for gastroesophageal cancer have not reported the rate of VTE during treatment or during the perioperative period.^25,26 In a small phase II trial, a retrospective analysis, and a brief communication, reported rates of VTE for localized gastroesophageal cancer were 6.1% to 16% in patients treated with chemoradiotherapy and 30% in patients treated with preoperative biochemotherapy.^18,27,28

The rates of VTE in adjuvant and preoperative therapy trials for most other GI cancers are equally under-reported, with the exception of colon cancer. In the MOSAIC trial, which investigated two different chemotherapy regimens for adjuvant therapy in stage II and III colorectal cancer, the rate of VTE among patients who received at least one cycle of chemotherapy was 6.1%.²⁹ The effort at reporting VTE has been duplicated in other recent phase III trials in adjuvant therapy for colorectal cancer.^30,31

Thus, the risk of VTE during adjuvant therapy for GI cancers currently is not well defined. Trial investigators should be encouraged to report VTE as part of their original publication or in publications of correlative studies. It would be beneficial for younger investigators to seize an opportunity to fill the gaps in knowledge about VTE in GI cancers by adopting such reporting as a routine part of the reporting of trial results.

RISK FACTORS

The risk factors for the development of VTE can be divided into intrinsic and extrinsic factors. More data on risk factors are available for pancreatic cancer than for other GI cancers. In a study by Blom et al in 202 patients with pancreatic cancer, patients with tumors of the corpus and cauda of the pancreas had a 2- to 3-fold increased risk of VTE compared with tumors located in the caput.³² In addition, presence of distant metastasis increased the risk of VTE 2-fold (hazard ratio [HR] 1.9, 95% confidence interval [CI] 0.7–5.1). Other intrinsic tumorrelated risk factors may include high levels of tissue factor expression, platelet aggregation induced by tumor cells, high levels of plasma plasminogen activator inhibitor-1, and elevated levels of other proteins activated in the coagulation cascade.^32–35

With regard to extrinsic factors, Blom et al also determined that receiving chemotherapy (adjusted HR 4.8) and the 30-day postoperative period (adjusted HR 4.5) were associated with increased risk for VTE. Patients receiving radiotherapy and those with mucinous adenocarcinomas of the pancreas did not appear to be at greater risk for VTE. Other series have confirmed that the location of the primary pancreatic tumor affects the risk of VTE, while challenging the notion that mucinous adenocarcinomas are not associated with increased risk.^36,37

It is expected that surgery and the presence of metastatic disease would increase the risk of VTE in patients with pancreatic cancer, since both conditions contribute to Virchow’s classic triad of venous stasis, vessel wall damage, and an increased hypercoaguable state. Surgical trauma can lead to vessel wall damage while the presence of metastatic disease can lead to venous stasis due to vessel obstruction from tumor or adenopathy or decreased patient mobility. However, it is not clear why radiation therapy does not appear to increase the risk of VTE to the same degree as chemotherapy. One explanation could be the selection bias of patients for radiation therapy; patients with lower disease burden and better performance status are selected for aggressive local therapy, and these factors might be associated with lower overall risk for VTE.

Fewer data are available on intrinsic and extrinsic risk factors for VTE in other GI cancers. However, it is likely that risk factors reported for pancreatic tumors and other solid tumors are applicable to other GI cancers, as well. These variables include treatment with cytotoxic chemotherapy, angiogenesis inhibitors, erythropoietin stimulating agents and endocrine/metabolic agents, presence of metastatic disease, number of medical comorbidities, tumor histology, presence of central venous catheters, and compression or obstruction of vessels by tumors or adenopathy (Table 2).^38–45

Table 2.

Reported risk factors for development of VTE in cancer³⁸–⁴⁵

Clinical Stage   Advanced > localized   Location of Primary Tumor

Histology   Adenocarcinoma > squamous-cell carcinoma   Mucinous > nonmucinous

Chemotherapy Agents   Cisplatin > oxaliplatin   Irinotecan

Vascular Epithelial Growth Factor Inhibitors   Bevacizumab (arterial thrombotic events)

Erythropoiesis-Stimulating Agents   Darbepoetin   Erythropoetin

Radiotherapy

Surgery

Medical Comorbidities   Number of conditions

Procoagulant Factors   Tissue factor   Plasminogen activator inhibitor-1   Cancer procoagulant

Hereditary Risk Factors   Factor V Leiden   Prothrombin 2021A mutations

MORBIDITY AND MORTALITY

VTE is associated with considerable morbidity. Patients who develop clinically relevant DVT may suffer from many direct adverse effects of thromboembolism, such as peripheral edema, pain, chronic venous insufficiency, and pulmonary hypertension. In addition, anticoagulation therapy for thromboembolism can be a significant burden. The development of VTE in patients with advanced cancer complicates symptom management and the already daunting challenge of maintaining a patient’s quality of life. The morbidity of VTE further underscores the necessity to dedicate additional research not only toward prevention and management of VTE, but also toward maintaining or improving quality of life in patients with concurrent malignancy and thromboembolism.

Recognition of the effect of VTE on mortality in patients with GI cancers is just now emerging. Tetzlaff et al reported that patients with advanced gastric cancer treated in clinical trials had a significantly shorter median survival if they developed VTE before or during protocol chemotherapy compared with patients who never developed VTE (3.9 months vs. 8.7 months, P = .007).⁴⁶ In a study in 198 patients with localized gastroesophageal cancer, these investigators also found that patients developing VTE before or during chemoradiotherapy had a significantly decreased median survival compared with patients without VTE (17.7 months vs. 32 months, P = .014). The development of VTE was an independent prognostic factor for overall survival in multivariate analysis.¹⁸

The negative effect of VTE on survival has also been demonstrated in colorectal cancer and pancreatic cancer. In a retrospective analysis of two merged databases in California including 68,142 colorectal cancer patients, it was shown that VTE was a significant predictor of death within 1 year of cancer diagnosis for patients with local and regional stage colorectal cancer (HR 1.5 for both).³⁸ In what appears to be the first randomized trial to report the effect of VTE on survival in colorectal cancer, Mandalà et al found that the development of VTE in 203 patients with metastatic disease had a negative effect on survival.⁴⁷ This finding remained significant after adjusting for age, disease site, and treatment schedule in multivariate analysis (HR 1.6, 95% CI 1.0–2.5).

In 227 patients with unresectable pancreatic cancer, Mandalà et al found that VTE during treatment was associated with a decrease in progression-free survival (HR 2.59, 95% CI 1.69–3.97) and overall survival (HR 1.64, 95% CI 1.04–2.58).⁴⁸ VTE during chemotherapy remained a significant prognostic factor after multivariate analysis. A second series reported by Zawin et al supports the observation that VTE is a poor prognostic factor.⁹ In this series, a high rate of VTE (71% overall) was found among 21 chemotherapy-naïve patients with advanced pancreatic cancer enrolled in a phase II trial; median survival was 8 months for patients with VTE compared with 21 months for patients without VTE.

Two series in gastroesophageal cancer did not show a detrimental effect of VTE on survival.^17,49 It should be noted that one study, in which 12 (25%) of 47 patients developed VTE, was not powered to detect a difference in survival.¹⁷ The second series⁴⁹ included patients with catheter-related thrombosis in the overall rate of VTE, which may have limited the ability of the investigators to detect a detrimental effect on survival. It could be the case that VTE that develops as a result of malignancy is a more accurate manifestation of aggressive tumor biology and thus associated with increased mortality, whereas VTE related to central venous catheters might be a manifestation of vascular trauma and endothelial reactivity that does not reflect tumor biology.

The role that VTE plays in patient survival merits further attention. Available data suggests that VTE is generally a poor prognostic factor for patients with cancer. It is unclear, however, whether the effect of VTE on survival is a direct reflection of aggressive tumor biology or rather related to the development and complications of VTE. If the latter is true, one strategy that might improve outcomes is selective anticoagulation for patients thought to be at increased risk for the development of VTE. Alternatively, if the development of VTE represents aggressive tumor biology, it may be possible to target anticancer therapy at the up-regulated pathway that leads to hypercoagulation and thrombosis—eg, with heparins. In animal models, heparins have been shown to inhibit cancer cell adhesion, proliferation, migration, and invasion, providing a rationale for further investigation of their use in preventing VTE.⁵⁰

ANTICOAGULATION AND SURVIVAL

Anticoagulation is recommended for patients with cancer and thromboembolism for the secondary prevention of recurrent thrombosis. In patients with cancer, anticoagulation with low molecular weight heparin (LMWH), unfractionated heparin, or vitamin K antagonists is often prescribed. Initial studies examining anticoagulation in patients with cancer focused on recurrent thrombosis and bleeding risks as end points. However, post hoc and subgroup analyses and meta-analyses from these trials have suggested that anticoagulation may have a beneficial effect on survival in patients with cancer and a history of thrombosis.^51–55 These findings remain controversial, though, since the reports of benefit involve heterogeneous cohorts of patients with differing primary tumors, stage of disease, and current treatment.

Several prospective trials have examined the effect of anticoagulation on survival in patients without a history of thrombosis. In a study by Klerk et al, 302 patients with advanced solid tumors were randomized to 6 weeks of treatment with daily nadroparin or placebo.⁵⁶ Patients were allowed to receive concomitant anticancer therapy at the discretion of the treating physician. The median survival was significantly longer in the nadroparin arm (8.0 months) compared with the placebo arm (6.6 months) (HR 0.75, 95% CI 0.59–0.96, P = .02). The survival difference remained significant after adjusting for performance status, histology, concomitant treatment, and life expectancy.

In contrast, the Fragmin Advanced Malignancy Outcome Study (FAMOUS trial), a randomized, double-blind, multicenter trial with the primary end point of mortality at 1 year, found no survival benefit with anticoagulation treatment.⁵⁷ In the FAMOUS trial, 385 patients with advanced solid tumors and no history of thrombosis received daily injections of dalteparin (5,000 IU) or placebo. There was no difference in survival at 1 year from randomization between the dalteparin group and the placebo group (10.8 months vs. 9.14 months, P = .19). In a post-hoc analysis, dalteparin improved median survival by 19.2 months in patients with a good prognosis and who had survived greater than 17 months from randomization compared with placebo patients in the same category. However, this analysis was not defined a priori and represented only 26% of the intention-to-treat population.

In a smaller phase III trial, 141 patients with advanced cancer were randomized to standard clinical care plus LMWH or standard care alone.⁵⁸ Median survival did not differ between the two groups in the intention- to-treat population, in the subgroup of patients with a favorable prognosis (patients that survived longer than 6 months), or in subgroups based on disease site. LMWH did not increase the rate of toxic effects for patients nor was it able to improve patients’ quality of life. A meta-analysis was performed to determine the potential effect of LMWH on survival in patients with advanced cancer, including the three trials discussed above plus a trial reported by Altinbas et al.^59,60 The meta-analysis suggested that LMWH plus standard therapy improved survival at 1 and 2 years with odds ratios of 0.70 (95% CI 0.49–1.00, P = .05) and 0.57 (95% CI 0.42–0.84, P = .03), respectively.⁶⁰ In addition, LMWH was not associated with an increased risk of bleeding.

The effect of anticoagulation on survival has been examined according to specific disease sites, as well. In a small randomized study in small-cell lung cancer, 84 patients received chemotherapy alone (cyclophosphamide/epirubicin/vincristine) or the same chemotherapy regimen with dalteparin (5,000 IU daily).⁵⁹ The duration of anticoagulation was limited to the 18 weeks of chemotherapy. At a median follow up of 10 months, there was a significant difference in median survival in favor of the patients treated with chemotherapy plus dalteparin vs. chemotherapy alone (10 months vs. 6 months, P = .01). Progression-free survival was also significantly longer in patients treated with dalteparin, and a nonsignificant trend in improvement in response rate was reported. The findings of this study are similar to other small studies in lung cancer, suggesting a benefit for treatment with heparin or coumarin derivatives.^61–63

At this time, insufficient evidence exists to support initiation of anticoagulation therapy in the absence of thromboembolism. Still, continued research is warranted in the subgroups of patients in which a suggestion of benefit has emerged. Initiating well-designed trials in locally advanced cancers of specific sites would be a rational approach to defining the potential utility of anticoagulation therapy, and such trials should be of interest to both patients and researchers.

The potential benefit of anticoagulation on survival in GI cancers cannot be defined from the data discussed above, since only a minority of the patients included in those trials had GI cancers. It is thus fortunate that a number of studies are actively recruiting patients to help answer this important clinical question (Table 3). Preliminary safety results have been presented from randomized phase II and III trials examining nadroparin, enoxaparin, or dalteparin in combination with chemotherapy in GI cancers. Three trials in pancreatic cancer have reported that LMWH in combination with gemcitabine was safe, but efficacy data are not yet mature.^64–66 Results from these trials and others in solid tumors are pending.

Table 3.

Ongoing and completed trials of low molecular weight heparin in gastrointestinal cancers

Study Identifier*	Patient population	Treatment	Trial status	Phase	Primary end point
NCT00718354	Gastric cancer	Standard CT ± enoxaparin	Recruiting	IIIb	Overall survival
NCT00312013	Pancreatic cancer Prostate cancer Lung cancer	Standard CT ± nadroparin	Recruiting	III	Death due to all causes at study end
NCT00003674	Colorectal cancer Breast cancer Lung cancer Prostate cancer	Standard CT ± dalteparin	Completed	III	Overall survival

Abbreviations: CT=chemotherapy.

^*www.clinicaltrials.gov

CONCLUSION

A significant line of research needs to be dedicated to investigation of thromboembolic phenomena in GI cancers. The effect of VTE on survival continues to be clarified. As well, the role of anticoagulation on the survival of patients with GI cancers also remains to be defined. Opportunities for additional research include investigation of methods to identify patients at risk for the development of VTE, as well as investigation of strategies and therapeutic interventions to reduce the morbidity and mortality of VTE. In addition, more convenient anticoagulation therapy schedules (daily/weekly) and routes (oral) are needed to improve the management of VTE while minimizing toxicity and reducing monitoring requirements.