Venous thromboembolism event (VTE) is a common and morbid complication in cancer patients. Patients with gastrointestinal cancers often suffer from symptomatic or incidental splanchnic vein thrombosis, impaired liver function and/or thrombocytopenia. These characteristics require a thorough risk/benefit evaluation for individual patients. Considering the risk factors for the development of VTE and bleeding events in addition to recent study results may be helpful for correct initiation of primary pharmacological prevention and treatment of cancer-associated thrombosis (CAT), preferably with low molecular weight heparins (LMWH). Whereas thromboprophylaxis is most often recommended in hospitalized surgical and non-surgical patients with malignancy, there is less agreement as to its duration. With regard to ambulatory cancer patients, the lack of robust data results in low grade recommendations against routine use of anticoagulant drugs. Anticoagulation with LMWH for the first months is the evidence-based treatment for acute CAT, but duration of secondary prevention and the drug of choice are unclear. Based on published guidelines and literature, this review will focus on prevention and treatment strategies of VTE in patients with gastrointestinal cancers.

Venous thromboembolism (VTE) commonly presents as deep vein thrombosis (DVT) or pulmonary embolism (PE), occasionally as thrombosis of the hepatic, portal, or splanchnic veins. Patients with cancer are at increased risk for VTE[1-4]. In addition to prothrombogenic cancer-related factors, the risk for symptomatic VTE is modulated by patient-specific and treatment-related factors (Table 1). Among the different subtypes of malignancy, some gastrointestinal cancers have a clinically relevant VTE incidence of more than 5% in the first year after cancer diagnosis (Figure 1), such as pancreatic (16%-22%), gastric (12%-17%) and colorectal (8%-12%) cancers[2,5-8]. Symptomatic VTE is not only associated with substantial morbidity and complications in clinical management, but has been shown to have a detrimental effect on cancer survival[9-11]. Acute and follow-up complications such as intestinal necrosis and esophageal bleeding due to portal hypertension may be life threatening if intra-abdominal veins are involved[12-15].

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Venous thromboembolic event-risk according to different cancer entities (modified from Wun et al[5] 2009). VTE-Incidence in the first year after cancer diagnosis (all stages) California Cancer Registry 1993-1999 (Patient Hospital Discharge Dataset). VTE: Venous thromboembolism event.

Furthermore, anticoagulation with a vitamin K antagonist (VKA), such as warfarin, is complicated by drug interactions and variable drug absorption due to the changing nutritional status of cancer patients[16]. In fact, therapeutic anticoagulation with VKA does not prevent VTE recurrences as effectively as in non-cancer patients, and major bleeding complications are also more than twice as common[17]. Acute anticoagulation and secondary prevention with low molecular weight heparin (LMWH) has been established as an effective treatment of VTE in cancer patients[18], but daily subcutaneous application may result in local complications such as hematoma or infections, thus negatively influencing quality of life.

Most research into the primary prevention and treatment of VTE included both cancer and non-cancer patients. Recommendations for patients with malignancies are therefore based on the sparse available evidence and the guidelines differ only marginally (Table 2)[19-27]. Most clinical studies focusing on VTE in cancer patients do not differentiate between different kinds of cancer, resulting in a lack of specific data concerning VTE in gastrointestinal cancers.

The risk for VTE is higher in cancer patients undergoing surgery without prophylaxis in comparison to non-cancer surgical patients[27,28]. Administration of LMWH or fondaparinux (FPX) in these patients significantly reduces the rate of VTE[29,30]. Dosages of LMWH in the range of 3000-5000 anti-FXa units per day are more effective than and as safe as lower doses. Patients undergoing major abdominal or pelvic surgery for malignancy remain at risk for VTE for up to five weeks after surgery. Thus, a prolonged pharmacologic thromboprophylaxis (LMWH or FPX once daily) should be considered, unless contraindicated because of high bleeding risk or active bleeding[31,32]. Prophylactic regimens begin 12-24 h pre- or 6-24 h postoperatively[29-32]. Based on these studies[31-33], and according to the different guidelines (Table 2), antithrombotic prophylaxis is recommended for a minimum of 7 d and up to 35 d postoperatively. FPX (2.5 mg/d) was found to be as effective and safe in preventing VTE after abdominal surgery as the LMWH dalteparin (5000 antiFXa units/d). A post-hoc analysis of 1407 (out of 2048) patients with cancer demonstrated a significant (39%) risk reduction of VTE[29]. Accordingly, the European (ESMO) guideline[20] recommends extended prophylaxis for all patients undergoing elective cancer surgery. In the American guidelines (ASCO, NCCN)[19,34], extended prophylaxis is only recommended in the presence of high thromboembolic risk factors such as residual or advanced cancer, aged 60 or older, obesity, previous history of VTE, duration of surgery longer than 2 h or prolonged postoperative immobilization. In patients with contraindications to pharmacological anticoagulant prevention strategies (e.g., increased risk of hemorrhage), the use of intermittent pneumatic compression devices or compression stockings is advised[19-27].

The recommendations for patients undergoing minimally invasive or laparoscopic surgery are even less evidence-based. In a recently published randomized study evaluating postoperative antithrombotic prophylaxis with LMWH for one vs four weeks in patients undergoing laparoscopic surgery for colorectal cancer, extended antithrombotic prophylaxis was safe and reduced the 3-mo risk of VTE by more than 90%, and that of proximal DVT by 50%, as compared to the one week regimen[35]. These data are consistent with the suggestion to follow the same recommendations regardless of whether a cancer patient is to undergo an open or laparoscopic surgical intervention[36,37].

Hospitalized patients with active cancer - a term not uniformly defined - were included in all published randomized clinical trials investigating the role of unfractionated heparin (UFH), LMWH or FPX for the prevention of VTE. Treatment of hospitalized patients for 6-14 d with low-dose enoxaparin (20 mg/d s.c.) was ineffective[43], whereas higher prophylactic doses of LMWH (enoxaparin 40 mg/d, or dalteparin 5000 anti FXa units/d)[43,44] or FPX (2.5 mg/d)[45] demonstrated superiority compared to placebo in the prevention of VTE with minimal or no increase in major bleeding events. Subgroup analyses failed to identify a subgroup which did not benefit from pharmacological thromboprophylaxis[46]. This was also true for a small subgroup (5%-15% of the study population) of cancer patients.

Subgroup analysis of 274 patients with active cancer from the CERTIFY study comparing UFH (3 × 5000 IU/d) with LMWH (certoparin 3000 anti FXa units/d) demonstrated a similar VTE risk (5.3% vs 4.1%) and similar rates of any (4.0 vs 3.9) or major (0.7% vs 0.5%) bleeding compared to the 2965 patients without cancer[47,48].

Based on three studies, extended prophylaxis with the anticoagulant drugs LMWH[49] or non-vitamin K oral anticoagulant drugs (NOACs)[50,51] cannot be recommended in medical patients. In the MAGELLAN trial[51], cancer patients (n = 592; 7.3%) had higher rates of VTE when prophylaxis with rivaroxaban was extended from 10 to 35 d.

Discussion is ongoing as to whether all cancer patients hospitalized for reasons such as infectious complications or complex chemotherapy regimens should generally receive medical thromboprophylaxis unless contraindicated by active bleeding or high bleeding risk[52-55]. According to the available guidelines (Table 2), routine thromboprophylaxis should at least be considered. Italian investigators assessing the risk of VTE in hospitalized medical patients confirmed cancer to be a major predisposing factor for VTE (PADUA prediction score), without differentiating between cancers[56]. However, the existence of active cancer on its own does not classify as a high VTE risk unless additional risk factors are present (Table 1).

The treatment of patients with active cancer and acute VTE with the standard treatment for non-cancer VTE patients (LMWH or FPX plus concurrent introduction of anticoagulation with VKA) resulted in a two- to three-fold increase of both re-thrombosis and bleeding[17,18]. Several randomized controlled trials compared prolonged application of LMWH to oral anticoagulation with VKA in patients with cancer-associated thrombosis (CAT) and demonstrated superior efficacy of LMWH with no increase in bleeding complications (Table 5), results confirmed by meta-analysis[18,19,80,81]. Based on these results, guidelines recommend a standard treatment of three to six months of weight-adjusted LMWH regardless of whether the patient suffers from gastrointestinal or another type of cancer (Table 2). The underlying malignancy, probably representing the most important risk factor for VTE, continues beyond six months after the initial VTE in most of these patients. Therefore, anticoagulation beyond this period may be warranted and is recommended in most guidelines (Table 2). Nevertheless, evidence for this recommendation is clearly weaker, as prospective randomized trials investigating type of anticoagulation, dosage of anticoagulant drugs or duration of prolonged anticoagulation have not been reported. Available data suggest that the pro-thrombotic risk correlates with disease progression due to increasing tumor load, decreasing performance status, progressive mobility reduction and increased use of palliative chemotherapy[82-84]. On the other hand, an increased bleeding risk, associated with progression of cancer, prolonged anticoagulation and prolonged chemotherapy needs to be considered in order to balance the benefit of preventing recurrent VTE against the risk of bleeding[83,84]. Furthermore, the impact of prolonged anticoagulation on the quality of life of cancer patients needs to be considered.

Although clinical guidelines (Table 2) currently recommend considering indefinite anticoagulation in patients with advanced malignancy, they do not recommend a specific anticoagulant drug due to the limited evidence available.

A number of new oral anticoagulants (NOAC) have been introduced and licensed for prevention and treatment of VTE in recent years[94]. These drugs work by direct inhibition of active coagulation factors direct oral anticoagulants (DOAC), namely factor IIa/thrombin (dabigatran) or factor Xa (apixaban, edoxaban, rivaroxaban). The term non-vitamin K oral anticoagulant was created in order to maintain the acronym NOAC. Although these drugs have been successfully tested and licensed in trials investigating NOACs in the post-operative setting of elective hip or knee replacement, drug development for other surgical patients was never initiated and the programs for medical patients were stopped early[50,51]. A placebo-controlled dose-finding phase II trial for primary prophylaxis of VTE in ambulatory cancer patients undergoing chemotherapy showed relevant activity[95], but this indication has not yet been further developed. There is therefore no evidence for the use of NOACs for primary prevention in hospitalized or ambulatory cancer patients undergoing surgery or any other kind of anticancer therapy.

The four drugs mentioned above have all been compared to VKA (warfarin) in randomized phase III trials for the treatment of DVT, PE and atrial fibrillation[96-101]. All of these trials demonstrated that NOACs are at least as effective and safe as VKA. As confirmed by meta-analysis, the study data for these NOACs showed a clinically relevant reduction in bleeding, most pronounced for intracerebral bleeding events[102]. Depending on the inclusion and exclusion criteria of the individual studies and on the definition of “patients with active cancer”, all six trials included cancer patients, a subgroup adding up to 2.5% of more than 25000 patients[102]. Meta-analysis of these patients suggests a potential role for NOACs in patients with CAT, with a similar risk/benefit relationship as demonstrated in non-cancer patients[76]. There are major drawbacks to the use of NOACs for this indication, however. First of all, standard therapy of CAT is LMWH for several months and not LMWH followed by VKA - the standard treatment arm in the VTE trials. Secondly, “patients with active cancer” included in the phase III trials demonstrate a three-month-mortality of less than 15%[103-105], whereas those included in the two pivotal CAT trials (CLOT, CATCH) recruiting cancer patients ten years apart, have a six-month-mortality of 40%[80,81] (Table 5). Obviously these patient groups differ to a great extent. Thirdly, despite having less interactions with other drugs compared to VKA, NOAC still bear a largely unclear risk of interactions with cytotoxic or targeted drugs. A CLOT- or CATCH-like head to head comparison of NOACs vs LMWH, the current standard of care, is eagerly awaited, especially as the oral application of NOACs might provide an improvement in quality of life compared to s.c. administered LMWH. Unless study data are available, NOACs are not to be considered the first choice in patients with acute CAT. As an alternative to LMWH or VKA (Table 2), anticoagulant treatment with NOACs beyond six months may be reasonable in many CAT patients, as there are no studies defining the optimal drug in this period.

The risk for VTE is clearly elevated in patients with (gastrointestinal) cancer. This risk is highest for patients with pancreatic cancer and those receiving anti-cancer therapies. VTE is the second leading cause of death in patients with cancer, and mortality is increased among patients with CAT[10,106]. Whereas available guidelines usually refer to VTE in cancer patients without differentiating between types of cancer, those with gastrointestinal cancers are more likely to present with additional problems such as hepatopathy-associated low platelet counts and/or prolonged prothrombin times. Furthermore, symptomatic or incidental VTEs of the visceral veins may occur more often. Despite limited data, prophylactic anticoagulation must be endorsed in most hospitalized patients with malignancies. In ambulatory patients undergoing chemotherapy, an assessment of individual prothrombotic and prohemorrhagic factors may help transfer the beneficial effect of pharmacologic prophylaxis demonstrated in a number of trials to those patients with the highest VTE risk, in particular those suffering from pancreatic cancer. Treatment recommendations for CAT with LMWH have been reconfirmed by recent evidence. Further progress will help clarify the risk/benefit relationship of NOACs in this field, identifying economic and quality of life aspects as well (Table 6).

The authors gratefully acknowledge the help of Mrs. Sue Travis to improve the English style.