Several studies consistently suggested that positivity of D-dimer at the time of warfarin discontinuation or soon after its interruption helps identify patients at a higher risk of developing recurrent VTE (Palareti et al. 2002, 2003; Eichinger et al. 2003; Verhovsek et al. 2008). These findings have recently been confirmed by a meta-analysis of available investigations (Bruinstroop et al. 2009) (Fig. 3).

Another post-baseline factor potentially associated with an increased risk of recurrent VTE is the early development of post-thrombotic manifestations (Stain et al. 2005).

In a few cohort studies, the ultrasound persistence of residual thrombosis after an episode of proximal DVT was found to be an independent risk factor for recurrent VTE (Prandoni et al. 2002; Piovella et al. 2002) (Fig. 4). Two recent meta-analyses of available investigations suggest that, whichever the method used for measuring the thrombotic mass, residual vein thrombosis is a powerful and independent predictor of recurrent VTE (Tan et al. 2011; Donadini et al. 2014). Of interest, residual vein thrombosis predicts not only the development of recurrent ipsilateral DVT, but also that of contralateral DVT and even of PE apparently not associated with DVT, suggesting that residual vein thrombosis is a marker of hypercoagulability (Prandoni et al. 2002; Piovella et al. 2002; Tan et al. 2011; Donadini et al. 2014).

In a recent prospective cohort study, residual vein thrombosis at 3 months after DVT (defined as the ultrasound incompressibility of at least 4 mm in the transverse section either in the popliteal or in the common femoral vein) was found to double the risk for subsequent recurrent VTE, post-thrombotic syndrome, arterial thrombotic events and cancer (Prandoni et al. 2015). The likelihood of residual vein thrombosis was higher in males, in patients with previous VTE and in those with extensive thrombosis. These findings suggest that a single assessment of residual thrombosis at 3 months can help risk stratify patients with proximal DVT and guide treatment decisions.

Whether the persistence of residual emboli after an episode of PE may predict the risk of recurrent PE as well is uncertain (Golpe et al. 2012). A recent prospective cohort study showed an unexpectedly high rate (85 %) of pulmonary artery recanalization in patients with PE treated with anticoagulants alone for 6 months, while in the remaining 15 % the thrombotic burden decreased by more than 80 % (Pesavento et al. 2014). These findings add to the current perception that residual thrombotic obstruction following PE is rare and unlikely to predict recurrent PE.

New categories of oral antithrombotic drugs are emerging. These drugs have the potential to simplify the treatment of patients with VTE by obviating the need for periodic laboratory monitoring and are associated with a favourable benefit-to-risk ratio. They include compounds that selectively inhibit factor Xa, such as rivaroxaban, apixaban and edoxaban, and compounds that selectively inhibit thrombin, such as dabigatran etexilate. The results of several randomized clinical trials suggest that these drugs have a favourable benefit-to-risk profile in the first 6–12 months after an episode of VTE in comparison with conventional treatment (Schulman et al. 2009, 2014; The Einstein Investigators 2010, 2012; Agnelli et al. 2013a; The Hokusai-VTE Investigators 2013). The value of long-term administration in preventing recurrent VTE has been evaluated in four studies performed in patients who had completed a 6–18-month period of conventional anticoagulation.

Two clinical trials were designed to assess the efficacy and safety of dabigatran for the extended treatment of VTE. In the RE-SONATE study, 1343 patients with unprovoked VTE who had completed 6–18 months of anticoagulant therapy were randomized to dabigatran (150 mg bid) or placebo for an additional period of 6 months (Schulman et al. 2013). A 92 % relative reduction in the relative risk for recurrent VTE was shown in favor of dabigatran, with a low risk for major bleeding (0.3 % vs 0). In the RE-MEDY study, 2856 patients at a higher risk of recurrent VTE who had been treated for 6 months with vitamin K antagonists for a first VTE were randomized to dabigatran (150 mg bid) or warfarin (INR 2–3) for the long-term prevention of recurrent VTE (Schulman et al. 2013). Dabigatran was shown to be non-inferior to vitamin K antagonists (1.8 % primary outcome events in dabigatran patients versus 1.3 % in warfarin patients). Major bleeding complications occurred in 0.9 % and 1.8 % of patients, respectively (reduction in the relative risk, 48 %).

The EINSTEIN Extension trial randomized 1197 patients with unprovoked VTE who had completed 6–12 months of anticoagulant therapy for an acute index VTE event to an additional 6–12 months of therapy with either rivaroxaban, 20 mg once daily, or placebo (The Einstein Investigators 2010). Recurrent symptomatic VTE events were recorded in 1.3 % of patients in the rivaroxaban group and 7.1 % of patients in the placebo group (relative risk reduction, 82 %). Major bleeding occurred in 4 (0.7 %) rivaroxaban-treated patients as compared to none in the placebo group.

The AMPLIFY-Extension was a 12-month randomized clinical trial where apixaban 2.5 mg and 5 mg bid were compared with placebo for extended treatment to prevent recurrent VTE in approximately 2500 patients with unprovoked VTE who had completed 6–12 months of treatment for DVT or PE (Agnelli et al. 2013b). Apixaban demonstrated superiority versus placebo in the reduction of the composite endpoint of symptomatic, recurrent VTE and death from any cause (p < 0.001). Apixaban was superior to placebo for the predefined secondary efficacy outcome of recurrent VTE and VTE-related death (8.8 % in the placebo group, compared with 1.7 % in both the apixaban 2.5 mg and 5 mg groups; p < 0.001). The rates of major bleeding were comparable for 2.5 mg (0.2 %), 5 mg (0.1 %), and placebo (0.5 %) treatment groups. The rate of the composite of major bleeding and clinically relevant non-major bleeding for the 5 mg treatment group (4.3 %) was higher than in the placebo group (2.7 %), while 2.5 mg treatment group (3.2 %) demonstrated similar rates to placebo.

Because of the potential to induce fewer bleeding complications, the new drugs may open new scenarios for decisions on the optimal duration of anticoagulation in patients with unprovoked VTE. However, these findings should be interpreted with caution for a number of reasons. Only selected patients were included in the clinical trials, thus no information is available for different patients categories, including carriers of major thrombophilias, patients with severe renal or hepatic failure, patients requiring anticoagulation for reasons other than VTE and so on. In only one study (the REMEDY) was the new anticoagulant (dabigatran) compared with warfarin in patients at a higher recurrence risk (Schulman et al. 2013), whereas in all the remaining the comparator was placebo, and the extent by which recurrent VTE was reduced over placebo did not exceed that achieved by vitamin K antagonists in studies adopting a similar study design (Boutitie et al. 2011). In only one study (the AMPLIFY-Extension) was there a head-to-head comparison between the conventional and a lower dose of the index compound (apixaban), whereas in all the remaining the initial dose was tested unchanged throughout the whole period of anticoagulation (Agnelli et al. 2013b). As the duration of these studies did not surpass 6–12 months, the persistence of a favourable benefit-to-risk ratio beyond this period cannot be anticipated. The results of the AMPLIFY-Extension study suggest that halving the initial dose after the first 6 months in patients with unprovoked VTE is likely to preserve protection against recurrent VTE while further reducing the haemorrhagic risk (Agnelli et al. 2013b).