Assessing Recurrence Risk Following Acute Venous Thromboembolism

Use of Algorithms

Daniela Poli; Gualtiero Palareti


Curr Opin Pulm Med. 2013;19(5):407-412. 

In This Article


With the aim to help clinicians in the difficult management of patients with a first unprovoked VTE event, in recent years three different algorithms have been proposed to identify patients who need long-term thromboprophylaxis and those in whom a short course of anticoagulation is sufficient.[9,10,11] The algorithms have been developed on the basis of analysis of clinical characteristics of patients, results of laboratory testing and of ultrasonography examination.

In 2008, Rodger et al.[9] proposed the HERDOO2 model, deriving data from a Canadian multicenter prospective cohort study that included 646 patients with a first unprovoked VTE event followed up for 18 months after discontinuing oral anticoagulation therapy ( Table 1 ). An annual risk of VTE recurrence of 9.3% [95% confidence interval (CI) 7.7–11.3%) was recorded in the whole cohort, and of 13.7% in men. On the basis of these results, the authors recommend a long-term anticoagulant treatment as always indicated for all men. With regard to women, the individual risk was stratified according to a score derived from the presence of specific risk factors: women with 0 or 1 score had an annual risk of recurrence of 1.6% (95% CI 0.3–4.6%), whereas those who had a score of 2 or more had an annual risk of 14.1% (95% CI 10.9–17.3%) and deserved, according to the authors, an indefinite anticoagulation.

In 2010, Eichinger et al.[10] published the Vienna prediction model that was validated on a prospective cohort study of 929 patients with a first unprovoked VTE. A total of 176 patients (18.9%) had recurrent VTE. The authors found that the risk of recurrence was higher in men [hazard ratio (HR) vs. women 1.90; 95% CI 1.31–2.75], if DVT was proximal (HR vs. distal 2.08; 95% CI 1.16–3.74), in cases of presentation as pulmonary embolism (HR vs. distal DVT 2.60; 95% CI 1.49–4.53), and when D-dimer levels, measured after discontinuation of anticoagulation, were increased (HR per doubling 1.27; 95% CI 1.08–1.51). Using these data, they developed a nomogram to compute the risk of recurrence ( Table 2 ).

The third algorithm has been recently published by Tosetto et al.[11] ( Table 3 ). The proposed prediction model is named DASH and includes the following factors: D-dimer level measured 1 month after anticoagulation withdrawal, young Age, male Sex, and Hormonal therapy associated with the index VTE event. The model was elaborated on the basis of data of a patient-level meta-analysis of 1818 patients with unprovoked VTE treated for at least 3 months with a VKA, with 239 recurrent VTE events. The DASH score varies between -2 and 4. The annualized rates of recurrence in relation to the scores were: 3.1% (95% CI 2.3–3.9) for scores of 1 or less; 16.4% (95% CI 4.8–7.9) for a score of 2, and 12.3% (95% CI 9.9–14.7) for scores of 3 or more. The DASH model considers at low risk of recurrence the patients who have a score of 1 or less; it has a satisfactory predictive capability (receivier operating curve area = 0.71). The authors conclude that a lifelong anticoagulation may be avoided in low-risk patients with unprovoked VTE.