Page 22 - 2019_07 resto del Mondo-web

1300
Editorials
Duration of anticoagulant treatment for unprovoked deep-vein thrombosis – is prolonged long enough?
Menno V. Huisman and Frederikus A. Klok
Department of Thrombosis and Hemostasis, Leiden University Medical Center, the Netherlands E-mail: MENNO V HUISMAN - m.v.huisman@lumc.nl
doi:10.3324/haematol.2019.218909
Patients presenting with deep-vein thrombosis or pul- monary embolism are often considered together as having one clinical syndrome, called venous throm- boembolism. From a pathophysiological perspective this makes sense, as both entities have their origin in the veins, most often the deep leg veins.1 In line with this, recent trials evaluating the direct oral anticoagulants have included both patients presenting with proximal deep-vein thrombosis (deep-vein thrombosis in the popliteal vein or more proxi- mal leg veins) and pulmonary embolism.2-5 However, look- ing at the natural history, there are some important differ- ences between patients with deep-vein thrombosis and those with pulmonary embolism. First, in a meta-analysis, patients who initially presented with deep-vein thrombosis had more deep-vein thrombosis as recurrent disease (79%) than pulmonary embolism (21%); patients with initial pul- monary embolism, had more recurrent pulmonary embolism (81%) than deep-vein thrombosis (19%).6 Second, whereas the rate of recurrent venous thromboem- bolism after acute pulmonary embolism closely resembles that of recurrent venous thromboembolism after a deep- vein thrombosis, the case fatality rate of recurrent deep- vein thrombosis or pulmonary embolism, after stopping anticoagulant treatment, in patients who presented with pulmonary embolism (5.7% to 12.3%) seemed to be higher than that of patients who presented with deep-vein throm- bosis (3.8% to 8.5%).7 Of note, in a more recent meta- analysis focusing solely on patients with unprovoked venous thromboembolism, this difference was not replicat- ed.8 Even so, it seems logical after all to evaluate antithrom- botic treatment for pulmonary embolism and deep-vein thrombosis separately.
As patients with unprovoked venous thromboembolism have a high risk of recurrent venous thromboembolism after stopping anticoagulant treatment, guidelines advocate continuing treatment after the initial 3 to 6 months.9 The basis for this recommendation was laid in a triad of trials comparing 3 months of anticoagulant therapy with a longer period up to 12 months.10-12 It became clear from these trials that the protection against recurrent venous thromboem- bolism provided by the prolonged therapy came at the obvious cost of (major) bleeding during anticoagulant treat- ment with vitamin K antagonists. Furthermore, the benefit of extended therapy subsided as soon as the treatment was stopped.
In the randomized PADIS-PE study13 in patients with acute unprovoked pulmonary embolism, which was per- formed by the same investigators as the PADIS-DVT study, after an initial period of 6 months of anticoagulation, an additional 18 months of warfarin therapy was compared to placebo; the benefit of the extended therapy was lost dur- ing a 2-year follow-up period after discontinuing anticoag- ulation. Subsequently, the same authors set out to evaluate
the benefit of extended therapy, using a similar design, in patients with unprovoked deep-vein thrombosis,14 for which they are to be commended. So, were the differences observed in the two earlier meta-analyses6,7 also apparent in this study?
The PADIS-DVT trial recruited 104 patients with acute unprovoked proximal deep-vein thrombosis without appar- ent major reversible risk factors for venous thromboem- bolism, including active cancer, within the 3 months pre- ceding the diagnosis of the deep-vein thrombosis. During the 18-month treatment period, recurrent deep-vein throm- bosis occurred in none of the 50 patients in the warfarin- group and in 16 out of 54 patients (cumulative risk, 29.6%) in the placebo group. During the first 12 months, this risk was greatest in the placebo group after anticoagulation dis- continuation (cumulative risk, 38.7% at 12 months). As the authors indicate, this risk, albeit comparable to the risk of recurrent venous thromboembolism in their PADIS-PE study, may well have been the result of the selection of study patients with a (very) high recurrence risk. Importantly, there is no mention of the location of the recurrent deep-vein thrombosis and how it was assessed, other than that ultrasonography was applied. The objective diagnosis of recurrent ipsilateral deep-vein thrombosis is particularly challenging, especially if prior ultrasound indi- cates residual thrombosis.15 Of note, in the PADIS-DVT study, more than 50% of patients had residual clots after the index deep-vein thrombosis. This may have caused an overestimation of the deep-vein thrombosis recurrence rate. Magnetic resonance direct thrombus imaging, a tech- nique that can differentiate acute from chronic thrombosis, might have led to an overall lower, more realistic estimate of the cumulative incidence of recurrent deep-vein throm- bosis.16
The advantage of continued anticoagulant treatment was lost upon discontinuation of the anticoagulation; during the entire 42-month study period, the composite outcome occurred in 14 patients in the warfarin group (cumulative risk, 36.8%) and 17 patients in the placebo-group (cumula- tive risk, 31.5%) (hazard ratio, 0.72; 95% confidence inter- val: 0.35-1.46). When the data from the PADIS-PE and PADIS-DVT studies were pooled, a non-significant hazard ratio of 0.68 indicated a potential overall benefit of prolong- ing anticoagulant treatment.14
The PADIS-DVT study has some limitations, of which, as noted by the authors, the most prominent one was the small sample size, due to the slow recruitment of patients and the study’s premature discontinuation. A remarkable finding was the low major bleeding rate over the whole 42- month study period, with only one major bleed in the war- farin group after treatment cessation (2.0%). This low rate may have been the result of patient selection and small sample size although in the PADIS-PE study, the major
haematologica | 2019; 104(7)