Skip to content
View commentaries on this research

Please note that this summary was posted more than 5 years ago. More recent research findings may have been published.

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.

A new prediction tool used following a spontaneous blood clot in the veins of the leg can identify people at risk of a further blood clot. This may help clinicians decide who should continue longer term treatment after initial treatment for the first clot has ended.

An anticoagulant such as warfarin is usually recommended for at least three months following a spontaneous or unprovoked blood clot, one that has occurred without any obvious cause. After this time, the risk of bleeding and cost of staying on anticoagulant medication needs to be weighed against the risk of a further blood clot. The researchers did not find a good way to help with the decision at this point, because one of the main predictors is a blood test which is unreliable when on anticoagulant treatment. They did develop a tool that was moderately accurate at identifying people at high risk of recurrence after they had stopped anticoagulation treatment for about a month.

The tool uses site of the blood clot, treatment duration, D-dimer (a clotting blood test), age and sex to determine the risk of a second blood clot.

Further testing is needed before it can be used in practice. In particular, more information is needed on bleeding risk for people on long term anticoagulant treatment.

Why was this study needed?

Every year in the UK, thousands of people have a blood clot in a vein (called venous thromboembolism). This includes clots in leg or arm veins (deep vein thrombosis) or in the lungs (pulmonary embolism). The usual treatment is with anticoagulants to prevent the clot spreading, most commonly with heparin to start with, then with tablets such as warfarin or with newer anticoagulant drugs that do not need as much monitoring.

At present, the minimum duration of anticoagulant treatment after having a clot in leg veins or lungs is three months. Longer treatment may be needed for people whose clot was unprovoked, which means lacking an obvious cause like prolonged immobility or following surgery. As the cause of the clot is unknown, the risk of recurrence is considered to be much higher than with known causes. A one-size-fits-all approach to treatment duration takes relatively little account of variation in other risks between individuals, such as older age or location of clots in the body and the researchers wanted to look at this further.

What did this study do?

The overall aim of this NIHR funded study was to develop and validate a model to predict risk of a second clot.

The study took place in three stages.

  1. A review of existing models to predict risk of clot recurrence and chance of other adverse outcomes after stopping anticoagulant treatment for a first, unprovoked clot in veins.
  2. The researchers then used the most promising predictors of the models they’d found, to develop and validate two new models.
  3. An economic model was developed to estimate the point at which – using this model – continued treatment is economically viable.

The development and validation of the new prediction model was carried out rigorously, and then tested or validated using different, large sets of patient data. The largest of these was the Spanish RIETE registry of records from 3,000 people since their first clot. Fewer data were available on costs, so we can be less certain of the cost effectiveness conclusions.

The D-dimer blood test is known to be inaccurate during anticoagulation treatment but becomes better after treatment stops (the lag time). The researchers developed and tested two models. One, the pre D-dimer model, with data collected immediately after cessation of treatment and the second the post D-dimer model, with data collected after about a month.

What did it find?

  • A post-D-dimer model was moderately effective at identifying who would have and who would not have a recurrent unprovoked blood clot (discrimination). It was based on sex, location of clot, age, D-dimer and time since stopping treatment. It had to be used about a month after stopping anticoagulant treatment, since the treatment affects the D-dimer test results. Using the c statistic, where 1 shows perfect discrimination and 0.5 shows no discrimination beyond chance, this model’s discrimination was moderate with an average c statistic of 0.69.
  • A pre-D-dimer model, used at the point of stopping anticoagulant treatment, based on just sex and location of clot, was poor at identifying who might have a recurrence or not. This model’s reported c statistic averaged 0.56 and (range 0.47 to 0.58), that is, not much better than chance.
  • The economic assessment estimated that the best trade-off between benefits and harms was to continue treatment with warfarin for those people with more than 8% risk of clot recurrence according to the post-D-dimer tool.

What does current guidance say on this issue?

The NICE guideline from 2012 on venous thromboembolic diseases recommends continuing anticoagulant treatment with Vitamin K antagonists (such as warfarin) for at least three months after having a deep vein thrombosis or pulmonary embolism. Extended treatment beyond three months is recommended for people with an unprovoked pulmonary embolism depending on their bleeding risk. NICE recommend that it should also be considered for people with blood clots in deep veins above the knee.

A British Committee for Standards in Haematology guideline published in 2011, states that anticoagulant treatment should be continued if the risk of bleeding is less than the risk of clot recurrence. They also recommend that long-term therapy is considered for people with unprovoked blood clots.

What are the implications?

The researchers were not able to find a reliable tool to help decide if someone should stop anticoagulation after three months while on treatment because one of the main predictors is the D-dimer blood test which is inaccurate during treatment.

However, the researchers have developed a new tool that can help identify people at high risk of clot recurrence who have stopped therapy. The tool combines readily available data plus a simple blood test so could easily be used in clinical practice to aid decision-making. It could provide reassurance that a future clot is unlikely or identify those at higher risk who may benefit from long-term therapy.

As this was a modelling study, using data already collected for other purposes, further research is needed on how the models would work in practice. It is unclear how many people would benefit from this approach and whether the long term use of this prediction tool, by providing a rational basis for stopping or continuing treatment, could reduce both rates of further clots and bleeding in this hard to manage group of people.


Citation and Funding

Ensor J, Riley RD, Jowett S et al. Prediction of risk of recurrence of venous thromboembolism following treatment for a first unprovoked venous thromboembolism: systematic review, prognostic model and clinical decision rule, and economic evaluation. Health Technology Assessment. 2016;20(12):1-190.

This project was funded by the National Institute for Health Research Health Technology Assessment Programme (project number 10/94/02).



Keeling D, Baglin T, Tait C, et al. Guidelines on oral anticoagulation with warfarin – fourth edition. Br J Haematol. 2011;154(3):311–24.

NHS Choices. Anticoagulant medicines [internet]. Leeds: NHS Choices; 2015.

NHS Choices. Blood clots [internet]. Leeds: NHS Choices; 2015.

NICE. Venous thromboembolic diseases: diagnosis, management and thrombophilia testing. CG144. London: National Institute for Health and Care Excellence; 2012.

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


NIHR Evidence is covered by the creative commons, CC-BY licence. Written content and infographics may be freely reproduced provided that suitable acknowledgement is made. Note, this license excludes comments and images made by third parties, audiovisual content, and linked content on other websites.

  • Share via:
  • Print article


An episode of venous thromboembolism was classified in this study as unprovoked where there was no history in the previous three months of any of the following risk factors:

  • major surgery
  • lower limb trauma
  • use of combined oral contraceptive pill or hormone replacement therapy
  • pregnancy
  • significant immobility
  • cancer

The final model specification had the following inputs:

  • Age
  • Sex (Males were more likely to have a future event)
  • Site of index event 9 More likely if the DVY was proximal or associated with a pulmonary embolus)
  • D-dimer
  • Lag time in days (the time between end of therapy and D-dimer testing)


Back to top