Skip to content
View commentaries on this research

This is a plain English summary of an original research article. The views expressed are those of the author(s) and reviewer(s) at the time of publication.

Unprovoked venous thromboembolism (VTE), including deep vein thrombosis in the leg and pulmonary embolism, are clots within veins that occur spontaneously in people without risk factors and are treated with anticoagulant drugs. If those drugs are stopped after three months or more, the risk of another clot appears to be on average 10% in the first year, 16% by two years, 25% by five years and 36% by 10 years.

This systematic review and meta-analysis of 18 studies included a total of 7,515 patients whose first thrombosis was unprovoked or brought about by minor factors.

The risk of recurrence is much lower where the first clot forms below the knee and men appear to have a greater risk than women, by about 40%.

The mortality over 10 years is about 1.5%, which is similar to the mortality from major bleeding when taking anticoagulants, in other studies. Considerations such as age, risk of bleeding, and other health problems will affect this balance of harms and benefits in individual patients, but these results provide a good starting point for the treatment discussion.

Why was this study needed?

Unprovoked venous thromboembolism (VTE) is a leading cause of death and disability in the UK and worldwide. Although it’s difficult to assess costs associated with deep vein thrombosis and pulmonary embolism accurately, the annual burden on the NHS has been estimated at almost £198 million.

We know that anticoagulant treatment following ‘unprovoked VTE’ (not associated with a major risk factor such as immobility or flying), is effective at preventing recurrence, but it can be difficult for clinicians and their patients to decide how long to continue treatment. The benefits of treatment are not maintained if anticoagulation is stopped at three to six months, but must be weighed against the risks of bleeding when anticoagulant treatment is continued indefinitely.

This meta-analysis includes some long-term and recently published studies to build our understanding of the risks of stopping anticoagulation for this group of patients.

What did this study do?

This systematic review and study-level meta-analysis looked at 18 studies (mainly trials) in developed countries with a total of 7,515 patients. Three studies followed patients for 10 years after stopping anticoagulation treatment; one study followed patients for 5 years, and the rest for up to 2 years.

The studies were used to estimate the risk of a second thrombosis event for patients who discontinue anticoagulation treatment for the first episode of unprovoked VTE.

There were differences between these non-UK studies, which may have weakened the conclusions. For example, there was some variation in the definition of ‘unprovoked VTE’ across the studies. Also, the results present only average risks within groups based on gender and extent of the original thrombosis, without accounting for age and other factors.

What did it find?

  • In patients with a first episode of unprovoked VTE who completed at least three months of anticoagulant treatment, the average risk of recurrent VTE was 10% in the first year after treatment, 16% at two years, 25% at five years, and 36% at 10 years, with 4% of recurrent VTE events resulting in death.
  • The cumulative risk of recurrent VTE at 10 years after discontinuation of anticoagulation was 41% in men and 29% in women with a first unprovoked VTE event.
  • Compared with patients with isolated pulmonary embolism, the rate of recurrent VTE was higher in patients with proximal deep vein thrombosis (rate ratio 1.4, 95% confidence interval 1.1 to 1.7). Over 10 years, men experienced an average 1.64% risk of fatal recurrent VTE. For women, there was a 1.16% risk.
  • The risk of recurrence after isolated distal thrombosis was only a fifth of the risk after either proximal vein thrombosis or a pulmonary embolism.
  • The risk of bleeding was not measured here, but the authors suggest an estimated 1.32% ten-year risk of fatal major bleeding in other studies if continuing anticoagulation.

What does current guidance say on this issue?

NICE recommends that anticoagulant treatment is continued in patients with unprovoked pulmonary embolism or deep vein thrombosis beyond three months, taking into account the risks and benefits.

The guidance advises clinicians to assess a patient’s individual risk of bleeding, and the likelihood of recurrent VTE, and to discuss this assessment with the patient, as well as providing written patient information about anticoagulant treatment.

What are the implications?

This study highlights the substantial risk for this group of patients of a recurrent dangerous blood clot. The risk of a second clot in the general population may be a little higher than in this selected group of trial participants, as may the risk of bleeding.

This study provides average risks by thrombosis site and gender, but other factors will modify that risk in individuals. The lower risk after distal thrombosis below the knee means that the potential harms of long-term anticoagulation may outweigh the benefits in many people.

These findings broadly support current NICE guidance in this area, while providing additional information for clinicians and guidelines developers.

Citation and Funding

Khan F, Rahman A, Carrier M et al. Long term risk of symptomatic recurrent venous thromboembolism after discontinuation of anticoagulant treatment for first unprovoked venous thromboembolism event: systematic review and meta-analysis BMJ. 2019;366:l4363.

This study was funded by the Canadian Institutes of Health Research.



NICE. Venous thromboembolic diseases: diagnosis, management and thrombophilia testing. CG144. London: National Institute for Health and Care Excellence; updated November 2015.

Thrombosis UK. Thrombosis statistics. Llandeilo: Thrombosis UK; 2016/17.

Produced by the University of Southampton and Bazian on behalf of NIHR through the NIHR Dissemination Centre


  • Share via:
  • Print article
Back to top