Cancer patients are often resistant to oral anticoagulant therapy while at the same time exhibiting a higher hemorrhagic risk. The best evidence of a higher risk of recurrence and clinically relevant hemorrhages while patients are receiving anticoagulation comes from a retrospective analysis of data from two large randomized clinical trials (Hutten et al. 2000) and two prospective cohort studies (Palareti et al. 2000; Prandoni et al. 2002). Hutten et al. extracted the rates of recurrent VTE and major bleedings in more than 1300 patients receiving at least 3 months of oral anticoagulant therapy for an acute episode of DVT (Hutten et al. 2000). The overall incidence of recurrent thrombosis in patients with cancer was 27.1 per 100 patient-years, versus 9.0 per 100 patient-years in those without cancer. The risk of major bleeding was 13.3 per 100 patient-years and 2.1 per 100 patients-years, respectively. Palareti et al. compared the outcome of anticoagulation courses in 95 cancer patients and 733 patients without malignancy (Palareti et al. 2000). Based on 744 patient-years of treatment and follow-up, there was a trend towards a higher rate of thrombotic complications in cancer patients (6.8 % vs 2.5 %; relative risk = 2.5). The rate of major bleeding was significantly higher in cancer patients (5.4 %) than in those without malignancy (0.9 %; relative risk = 6.0). We conducted a prospective cohort study in 842 consecutive patients with DVT who were administered conventional anticoagulation, of whom 181 were carriers of cancer (Prandoni et al. 2002). The 12 month cumulative incidence of recurrent thromboembolism in cancer patients was 20.7 % versus 6.8 % in patients without cancer, for an age-adjusted hazard ratio of 3.2 (95 % CI, 1.9–5.4) (Fig. 1). The 12 month cumulative incidence of major bleeding was 12.4 % in patients with cancer and 4.9 % in patients without cancer, for an age-adjusted hazard ratio of 2.2 (95 % CI, 1.2–4.1) (Fig. 2). In summary, cancer patients have a three to fourfold higher risk of recurrent VTE during anticoagulant therapy than cancer-free patients, very likely as a consequence of the release of cancer procoagulants that are not inhibited by conventional anticoagulation. This risk correlates with the extent of cancer (Prandoni et al. 2002). Recently, a stratification score has been developed and validated that has the potential to help clinicians predict the VTE recurrence risk and thus tailor treatment, improving clinical outcomes while minimizing costs (Table 3) (Louzada et al. 2012).

According to the results of three randomized clinical trials, LMWHs in full doses for the first month followed by a dose ranging from 50 to 100 % of the initial regimen have the potential to provide a more effective antithrombotic regimen in cancer patients with venous thrombosis than the conventional treatment and are not associated with an increased hemorrhagic risk (Lee et al. 2003; Meyer et al. 2002; Hull et al. 2006) (Fig. 3), even in patients with disseminated cancer such as those with liver or brain metastases (Monreal et al. 2004). In addition, LMWHs provide an anticoagulation that is easier to administer, more convenient and flexible, and not influenced by nutrition problems or liver impairment (Kearon et al. 2012). Thus, the long-term administration of LMWH should now be considered the treatment of choice in patients with metastatic disease and in those with conditions limiting the use of oral anticoagulants (Kearon et al. 2012; Lyman et al. 2015; Farge et al. 2013).

After discontinuation of antithrombotic treatment, cancer patients with venous thrombosis present a risk for recurrences that is almost twice as high as that observed in patients free from malignancies (Prandoni et al. 1996; Hansson et al. 2000; Heit et al. 2000). Among the factors associated with an increased risk of recurrent VTE after anticoagulation withdrawal are residual vein thrombosis, as determined by compression ultrasound 3 months after the index event (Donadini et al. 2014), and abnormal D-dimer values, measured on the day of drug suspension and 1 month later (Cosmi et al. 2005). In view of the persistently high risk of recurrent thrombotic events, prolongation of anticoagulation should be considered for as long as the malignant disorder is active provided that it is not contraindicated. For most patients, this translates into life-long anticoagulation. This decision should be frequently reassessed during patients’ follow-up.

The treatment of challenging situations has recently been addressed by the Subcommittee of the International Society of Thrombosis and Haemostasis (Carrier et al. 2013).

As new categories of drugs have emerged that have the potential to replace conventional treatment for the initial and long-term treatment VTE, major improvements are expected for the management of cancer patients with venous thrombosis. They include direct inhibitors of factor Xa (such as rivaroxaban, apixaban and edoxaban) and direct inhibitors of factor IIa (such as dabigatran etexilate). They possess several advantages over conventional drugs, including the inhibition of fibrin-bound Xa or thrombin, respectively, a dose response that is more predictable because there is no binding to plasma proteins, and a lack of potential to produce immune thrombocytopenia. As a consequence of their pharmacokinetic and pharmacodynamic properties, they can be administered orally, in fixed doses, without laboratory monitoring. Based on available information coming from well designed and conducted phase-3 randomized clinical trials, they possess a more favourable benefit-to-risk profile than the old compounds, make it possible to implement the treatment from the beginning, and cover the whole spectrum of clinical presentations, including severe pulmonary embolism (Schulman et al. 2009, 2013; The Einstein Investigators 2010, 2012; Agnelli et al. 2013a, b; The Hokusai-VTE Investigators 2013).

However, for the time being their use in patients with cancer requires caution. Indeed, only a small minority of patients with cancer (consistently around 5 %) were included in the abovementioned studies. Although based on the results of recent subgroups analyses (Prins et al. 2014) and meta-analyses of available studies (Vedovati et al. 2015; van der Hulle et al. 2014; van Es et al. 2014), the benefit-to-risk ratio of these new drugs for the treatment of cancer patients with VTE is encouraging, there is the need for a direct comparison with the LMWHs, which represent the standard of treatment in cancer-associated thrombosis (Kearon et al. 2012). Severe liver and renal dysfunctions, which contraindicate the use of all new oral compounds, are quite common in patients with cancer. There is uncertainty about the proper management of patients requiring emergency procedures and in those with thrombocytopenia. In conclusion, we believe that the novel anticoagulants need further investigations before routine usage in cancer patients.

Asymptomatic PE is a common finding in medical oncology due to the routine use of modern computed tomography scanners for cancer staging (Fig. 4). Although the clinical relevance of these incidental findings is unknown, based on the results of a few investigations conducted in recent years they are likely to impact on both the incidence of recurrent VTE and on the overall prognosis to the same degree as the symptomatic findings (Menapace et al. 2011; Sahut D’Izarn et al. 2012; Agnelli et al. 2014; Connolly et al. 2013; den Exter et al. 2011; Heidrich et al. 2009; O’Connell et al. 2010). Accordingly, the most recent international guidelines recommend the same initial and long-term anticoagulation as for comparable patients with symptomatic PE (Kearon et al. 2012).

However, there is still uncertainty about the optimal management of patients with the incidental detection of isolated subsegmental PE. Indeed, whether the outcome of these patients is comparable to that of patients with symptomatic involvement of subsegmental arterial vessels (den Exter et al. 2013) is unknown, as is the accuracy of detecting PE on CT-scans that were not specifically ordered to diagnose PE. Higher risk of false positive diagnosis compared to patients with suspected PE cannot be excluded. In a series of 70 patients diagnosed with subsegmental PE, this diagnosis was confirmed in only 51 % by a reviewing radiologist (Pena et al. 2012). PE may not be acute but chronic. In a series of 65 cases of untreated subsegmental PE reported so far, none of these patients developed recurrent VTE (Donato et al. 2010). Finally, there is uncertainty about the outcome of patients with incidental PE who receive anticoagulants. In a cohort study of 51 patients with incidental PE, 5 (9.8 %) patients developed major bleeding of which 2 cases were fatal (den Exter et al. 2011). Thus, a careful evaluation should be individually done rather than giving full-dose anticoagulation to all cancer patients with the occasional detection of isolated subsegmental PE.