Rationale and impact
- D-dimer testing
- Pulmonary embolism rule-out criteria (the PERC rule)
- Outpatient treatment for low-risk pulmonary embolism
- Anticoagulation treatment for suspected or confirmed deep vein thrombosis or pulmonary embolism
- Long-term anticoagulation for secondary prevention
- Inferior vena caval filters
- Investigations for cancer
These sections briefly explain why the committee made the recommendations and how they might affect practice. They link to details of the evidence and a full description of the committee's discussion.
The committee agreed that, if both laboratory-based and point-of-care D-dimer testing are immediately available, laboratory testing is preferable because it provides more rigorous quality assurance and greater certainty of diagnostic accuracy. However, if laboratory-based testing is not immediately available, the committee were in agreement that offering immediate point-of-care testing is more beneficial for patients than delaying diagnosis by waiting for laboratory testing. Although point-of-care tests are more expensive than laboratory tests, a cost-effectiveness analysis showed that the additional cost may be offset by faster results that reduce the need for additional GP time and unnecessary interim anticoagulation.
Evidence on fully quantitative point-of-care D-dimer tests for deep vein thrombosis (DVT) suggested that they are as accurate as laboratory tests and more accurate than qualitative or semi‑quantitative tests. There is little evidence on these tests for pulmonary embolism (PE) but the committee agreed that the evidence on DVT is applicable to PE because it is very unlikely that there is a biological reason that the accuracy of the tests would differ between these groups. The committee were aware that quantitative tests are more commonly used than qualitative or semi-quantitative tests. However, because the latter types of tests are still used in some services, they specified that point-of-care tests should be fully quantitative.
In people aged over 50, there was limited prospective evidence available for DVT and only retrospective evidence available for PE. This evidence suggested that adjusting D-dimer test thresholds for age improves the usefulness of these tests for ruling out venous thromboembolism (VTE) in this age group. The evidence also suggested that age adjustment does not reduce the accuracy of the tests in identifying VTE. The committee noted that adjusting test thresholds for age could be beneficial in reducing anxiety and unnecessary imaging for people with suspected DVT or PE. Although the evidence was not plentiful, the committee agreed that, taken together with the potential benefits, it was sufficient to support a recommendation suggesting age adjustment in D-dimer test thresholds for people aged over 50.
Services that do not currently provide quantitative point-of-care D-dimer tests may need to acquire new equipment and provide training on how to conduct and interpret the tests. It is uncertain what impact this would have on practice, because it is unclear what proportion of primary care centres already use point-of-care testing. Facilities for point-of-care testing are only needed if rapid laboratory testing is not available.
In people with signs or symptoms of PE, but in whom clinical suspicion of PE is low (the clinician estimates the likelihood of PE to be less than 15% based on the overall clinical impression and other diagnoses are feasible), there was some evidence showing that the PERC rule can accurately eliminate PE as a possible diagnosis. The committee agreed that using the PERC rule can reduce anxiety and avoid unnecessary D-dimer testing, imaging and interim anticoagulation treatment for people with a low probability of PE and none of the PERC criteria for PE. However, the evidence was limited so the committee agreed to recommend that the PERC rule be considered as part of initial assessment. The committee noted that the studies evaluating PERC all took place in emergency departments but they could see no reason why its use should be limited to this setting or why the diagnostic accuracy of PERC would differ in other settings.
The PERC rule is not widely used in current practice. This recommendation is expected to increase its use within a subgroup of people in whom clinical suspicion of PE is low and for whom discharge is being considered. Increased use of PERC can be expected to reduce the need for D-dimer testing and imaging for people with none of the PERC criteria for PE, leading to some reductions in waiting times in primary care and emergency departments. It will also help to avoid unnecessary anticoagulation treatment. However, the overall impact of this recommendation is not expected to be substantial because of the limited population it affects.
The committee noted that outpatient treatment for people with PE who have a low risk of poor outcomes is increasingly being used in settings such as ambulatory care units. There was limited evidence comparing outpatient with inpatient treatment for PE so the committee were unable to reach firm conclusions about the overall benefits and risks of outpatient treatment. However, no evidence showed that outpatient treatment is less effective or less safe than inpatient treatment for people with low-risk PE. The committee agreed that outpatient care offers substantial benefits for people with PE and for hospital services and should be considered for those with suspected or confirmed low-risk PE.
The committee emphasised the importance of clear arrangements for monitoring and follow-up to ensure that outpatients receive the same quality of care as inpatients. They noted that specialist services with expertise in thrombosis are not available at all times and agreed that it is important for people with VTE to know who they can contact if they need advice outside normal service hours.
Outpatient treatment for PE is common practice in many services with ambulatory care units. These recommendations might lead to the establishment of ambulatory care units in services that do not currently have them, and this will reduce hospital stays in those services.
There was no evidence specifically on interim anticoagulation treatment for suspected DVT or PE. However, the committee agreed that it is vital to start treatment if DVT or PE is suspected and diagnostic test results are delayed by more than 4 hours. They reasoned that anticoagulation treatments that are effective for confirmed DVT or PE are likely to be equally effective when used as interim treatment while awaiting a confirmed diagnosis. They also noted that continuing the same anticoagulant treatment after diagnosis offers benefits in terms of safety and convenience. However, they acknowledged that local protocols or availability of anticoagulants for suspected VTE may necessitate the use of different anticoagulants before and after diagnosis.
The committee agreed that when choosing an interim anticoagulant, clinicians should always take individual clinical circumstances into account, including whether the person is at an extreme of body weight, has PE with haemodynamic instability, renal impairment, active cancer or established triple positive antiphospholipid syndrome.
Evidence suggested that treatment with a direct-acting oral anticoagulant (DOAC) is less likely to result in bleeding complications than treatment with low molecular weight heparin (LMWH) and a vitamin K antagonist (VKA). Additionally, people taking a DOAC benefit by being able to have an oral treatment and avoid the frequent monitoring that is necessary with other types of anticoagulation treatment.
Within the DOACs, there was evidence showing that apixaban is the most cost‑effective option. because it results in the fewest bleeds. Rivaroxaban is the second most cost-effective option and only slightly less cost effective than apixaban. However, the committee had reservations about this evidence because the inclusion criteria setting out which patients took part in the studies were not the same in each study. In particular, the apixaban study did not include patients with provoked VTE unless it was caused by a persistent risk factor, so a larger proportion of patients in the apixaban study had unprovoked VTE compared with the rivaroxaban studies. This made it difficult to compare the results of the studies. Because of this, the committee were not confident that apixaban should be the only option for a DOAC and recommended a choice of apixaban or rivaroxaban. Sensitivity analyses were carried out varying the drug prices but these analyses did not change any of the conclusions from the economic model.
The committee recognised that apixaban or rivaroxaban might not be suitable for everyone, so they included options for treatment with LMWH followed by dabigatran or edoxaban, or LMWH with a VKA. The committee also made a recommendation for research on DOACs compared with each other and with other anticoagulants.
The evidence did not support a recommendation for fondaparinux. It showed that fondaparinux is more likely to result in bleeding and is less cost effective than other treatments. However, the committee decided not to make a recommendation precluding its use because they were aware that it may be needed in rare circumstances.
Unfractionated heparin (UFH) was associated with increased bleeding complications, greater recurrence rates of VTE and higher mortality rates than other treatments so the committee did not think it should be offered routinely. They recognised that it may be a suitable option for some people with VTE.
Anticoagulation treatment for DVT or PE in people at extremes of body weight (less than 50 kg or more than 120 kg)
Body weight can influence the absorption, distribution and elimination of anticoagulants, and their therapeutic effect can be altered at extremes of body weight. Because of this, and based on their knowledge and experience, the committee agreed that body weight should be taken into account and therapeutic monitoring considered when choosing anticoagulation for people whose weight is outside the range of 50 kg to 120 kg.
There was little evidence on the comparative effectiveness of different anticoagulants for people at extremes of body weight, and the evidence was limited to obesity defined as a body mass index of 30 kg/m2 and above rather than body weight. However, the committee noted that the summaries of product characteristics (SPCs) for several anticoagulants specify body weight rather than obesity and agreed that using the same criterion as the SPCs would make the recommendations clearer and easier to implement.
Because of the uncertainty about the most effective anticoagulant treatment for this group, the committee included a subgroup based on body weight in their recommendation for research on DOACs compared with each other and with other anticoagulants.
The committee agreed that intravenous UFH should be offered to people with PE and haemodynamic instability because the anticoagulant effect needs to be carefully controlled for these people. People with haemodynamic instability have poor peripheral circulation and because UFH is administered intravenously it allows for a more certain therapeutic effect. Additionally, the anticoagulant effect of UFH wears off relatively quickly if treatment needs to be stopped.
The committee did not review the evidence on thrombolytic therapy and the 2012 recommendation that it be considered for this population is unchanged.
Renal impairment increases the risk of anticoagulants accumulating in the body, which can increase bleeding risk. There was very limited evidence on anticoagulant treatment for VTE in people with renal impairment.
Based on their expertise and the SPC for each treatment, the committee agreed that LMWH, UFH or DOACs are suitable options to treat VTE in people with renal impairment. However, dabigatran is not an option for people with more severe renal impairment (estimated creatinine clearance 15 ml/min to 29 ml/min) based on its SPC. For people with estimated creatinine clearance less than 15 ml/min the main options are UFH or LMWH given either on their own or with a VKA until oral anticoagulation is established and in the therapeutic range. The committee emphasised the importance of following the SPCs and locally agreed protocols, and seeking advice from specialist colleagues or a multidisciplinary team to ensure correct dosing and monitoring.
There was very little evidence available on the duration of anticoagulation treatment for people with DVT or PE and active cancer. The committee agreed, based on the evidence and their experience, that anticoagulation treatment should continue for 3 to 6 months and then be reviewed. They noted that most of the evidence on VTE in people with active cancer looked at treatment over a period of 6 months, but agreed that some people need shorter treatment durations and it is good practice to determine the length of treatment on a case-by-case basis.
The effectiveness of DOACs compared with other anticoagulation treatments in people with active cancer has not been studied sufficiently to enable firm conclusions to be made. Evidence from studies in people without cancer may not be applicable because cancer could affect the action of these drugs. In studies that recruited only people with active cancer and VTE, rivaroxaban, edoxaban and LMWH were found to be similarly effective, although bleeding complications were more frequent with DOACs. These studies did not look at apixaban or dabigatran. In studies that looked at apixaban and dabigatran and in which a small number of people within the study population had active cancer, the effects of apixaban and dabigatran were similar in people with and without cancer. Economic evidence based on these studies showed apixaban to be the most cost-effective option, although the evidence for apixaban was based on a relatively small number of people.
The committee agreed that, if suitable, a DOAC should be considered to treat VTE in people with active cancer but, because of the limitations of the evidence, they could not be more specific about the choice of DOAC.
The committee noted the potential for interactions between anticoagulants and other drugs people with active cancer may be taking, such as chemotherapy drugs, and the possibility of an increased risk of bleeding with some types of tumours. The evidence suggested a higher rate of gastrointestinal and genitourinary bleeds in people with active cancer having treatment with a DOAC compared with those having LMWH. The committee agreed that DOACs may be unsuitable for people with tumours that are associated with an increased risk of these types of bleeds (such as people with gastrointestinal malignancies). However, they agreed that treatment decisions for people with active cancer need to be made on a case-by-case basis.
The committee made provision for people with active cancer if a DOAC is not suitable by including LMWH alone and LMWH with a VKA as alternatives. Although LMWH alone is commonly used in practice and is the only licensed option to treat VTE in active cancer, it is not cost effective compared with DOACs and reducing its use would be beneficial in conserving NHS resources. Sensitivity analyses were carried out varying the drug prices but these analyses did not change any of the conclusions from the economic model.
The committee recognised that there are circumstances in which LMWH is the most suitable treatment option and agreed that it could be considered when this is the case. They were aware that LMWH with a VKA is often impractical for people with active cancer because of difficulties with INR monitoring and maintaining INR within the therapeutic range. They agreed that it is less clinically effective than LMWH alone but is less costly and remains a suitable option in some circumstances.
The committee were aware of an MHRA safety alert warning of an increase in VTE recurrence in people with diagnosed triple positive antiphospholipid syndrome taking a DOAC compared with those taking LMWH and a VKA. Although people with antiphospholipid syndrome were not included in the evidence review, the committee agreed that it is important to include a recommendation highlighting the need to offer LMWH with a VKA to this group.
VTE can be difficult to treat in people who use intravenous drugs. They often have problems with access to medical care and adherence to prescribed treatments. There is a lack of good evidence on the comparative clinical and cost effectiveness of treatments and doses for VTE in this population. The committee made a recommendation for research with the aim of improving VTE treatment in people who use intravenous drugs.
The recommendations are expected to lead to increased use of DOACs, particularly apixaban and rivaroxaban, to treat suspected and confirmed VTE. This should reduce the need for resources to monitor INR, manage bleeding complications and administer parenteral anticoagulation. The recommendation to start anticoagulation treatment before blood test results are available may increase community prescribing of anticoagulation treatment. However, more use of DOACs may also increase the need for expensive reversal agents.
Current VTE management for people at extremes of body weight is not expected to change substantially.
For people with haemodynamically unstable PE, UFH is currently used in clinical practice and the recommendation is not expected to affect the frequency of its use in this group.
More people with renal impairment are likely to be offered a DOAC or LMWH, reducing the use of UFH. This can be expected to produce cost savings by increasing the number of people with renal impairment who can have outpatient care for VTE.
For people with active cancer, it is expected that there will be an increase in the use of DOACs and a concomitant decrease in the use of more expensive treatments such as LMWH alone. This will also reduce the amount of district nursing support needed to provide assistance with parenteral therapies.
For people with VTE and antiphospholipid syndrome, the use of DOACs is expected to decrease.
The committee agreed that the benefits of anticoagulation treatment become less certain over time, and that after 3 months (or 3 to 6 months in people with active cancer) treatment needs to be reviewed and a decision made about whether to continue or stop treatment. They agreed that, at this point, the aim of anticoagulation changes from treatment to reducing the risk of recurrence.
The committee noted that continuing anticoagulation treatment after 3 months is less beneficial for people who have had a provoked DVT or PE if the provoking factor is no longer present, because of the lower rate of recurrence compared with unprovoked DVT or PE.
For people with unprovoked DVT or PE, the benefits and risks of continuing anticoagulation treatment are less certain and the committee agreed that they need to be carefully balanced. However, for most people with a low bleeding risk, the committee agreed that the benefits of continuing anticoagulation treatment outweigh the risks.
The committee agreed that the tools currently available to predict the risk of recurrence of VTE or the risk of bleeding are not sufficiently accurate or validated to be used as the sole basis for a decision, and that using them in such a manner might result in incorrect predictions and subsequent harm to the person. However, they also agreed that, in certain circumstances, a clinical prediction tool can be a useful adjunct to discussion with people offered long-term anticoagulation treatment. For bleeding risk, evidence on the HAS-BLED score showed that it can identify people with unprovoked proximal DVT or PE who are at particularly high risk of major bleeding and might benefit from stopping anticoagulation. For VTE recurrence, there was very limited evidence, and that evidence suggested that these tools are not sufficiently accurate to be used in practice.
Because of the uncertainty in predicting VTE recurrence and the risk of major bleeding, the committee made recommendations for research to develop a new prediction tool for VTE recurrence and major bleeding combined and to compare this tool with clinical judgement.
The committee agreed that there are risks involved in switching anticoagulant treatment, particularly if there have been no adverse events with the current treatment. They also expressed concerns about convenience for people who are asked to switch from a DOAC with no monitoring to a VKA regimen with frequent monitoring, or problems with adherence if switching from a VKA to a DOAC. Based on these concerns and their clinical experience, the committee agreed that if treatment is continued beyond 3 months, the first option for most people should be to continue the current treatment if it is well tolerated.
Some evidence indicated that there are fewer major bleeds with apixaban than with rivaroxaban, dabigatran or a VKA. However, the committee were not entirely convinced by this evidence because the study of apixaban had stricter inclusion criteria, setting out which patients took part, than the other studies. Additionally, the studies recorded a very low number of major bleeds, leading to uncertainty about the effects of the different anticoagulation treatments on the likelihood of major bleeding. Apixaban was shown by the economic evidence to be the most cost-effective long-term treatment so the committee agreed that switching to apixaban should be considered as an option for people currently taking a DOAC other than apixaban. Sensitivity analyses were carried out varying the drug prices but these analyses did not change any of the conclusions from the economic model.
There was a lack of evidence on longer-term treatment for people with renal impairment, active cancer, antiphospholipid syndrome or extremes of body weight. Based on their clinical experience, the committee agreed that continuing the current treatment, if it is well tolerated, should be considered for people in these groups, taking into account their preferences and clinical situation.
For people who do not continue anticoagulation treatment, evidence showed that aspirin is better than no treatment at reducing DVT or PE recurrence for up to 2 years, although there was no difference in DVT or PE recurrence between aspirin and no treatment at 4 years. On balance, the committee agreed that aspirin can be considered as an option for people who wish to stop anticoagulation treatment, using a prophylactic dose based on current UK practice (75 mg daily or 150 mg daily).
To ensure that treatment is guided by the person's changing balance of benefits and risks, and changes in their preferences over time, the committee agreed that people taking long-term anticoagulation treatment or aspirin should have their risk of VTE recurrence, bleeding risk and general health reviewed at least once a year.
The recommendations to review anticoagulation treatment at 3 months (3 to 6 months for people with active cancer), and annually thereafter, reflect most current practice but may increase the number of appointments and clinician time needed in services that do not currently provide these reviews. Giving patients who stop anticoagulation treatment information on signs and symptoms and a point of contact should ensure that a comprehensive safety net is in place.
Discussions about the benefits and risks of stopping or continuing anticoagulation treatment may increase the time needed for consultations, particularly if a prediction tool is used as part of the decision-making process. Using HAS-BLED can be expected to reduce long-term anticoagulation treatment in people with a high risk of major bleeding.
The recommendations are not expected to change practice for people with renal impairment or extremes of body weight. For people with antiphospholipid syndrome, it is expected that DOAC use will decrease, with a concomitant increase in the use of LMWH with a VKA. For people with active cancer, DOAC use can be expected to increase, with LMWH on its own becoming a less common treatment. For people with none of these clinical features, greater use of DOACs, particularly apixaban, for long-term treatment can be expected to lower costs by reducing the need for clinical visits, INR monitoring and managing bleeding events. The use of aspirin may increase for people who decline long-term anticoagulation. This may reduce VTE recurrence in people who would otherwise not receive treatment.
There was little good evidence on inferior vena caval (IVC) filters. The evidence in a number of populations, including people about to have surgery, people with cancer, people with a high risk of having a subsequent PE and people with a high risk of a poor outcome from a subsequent PE, did not show a benefit from IVC filters. The committee therefore agreed that the use of IVC filters should be restricted to prospective clinical studies (including but not limited to prospective cohort studies and randomised controlled trials) unless anticoagulation is contraindicated or a PE has occurred during anticoagulation treatment. They made a recommendation for research to further investigate the effectiveness of IVC filters.
For people with proximal DVT or PE and a contraindication to anticoagulation treatment, a small amount of new evidence has become available since the 2012 guideline was published. This evidence did not show a clear difference in outcomes such as mortality and VTE recurrence between people who were given IVC filters and those who were not. One study found evidence of an increase in DVT occurrence (in people with an undefined initial VTE event) at 1 year in the group given IVC filters. However, based on their experience and knowledge of IVC filters, the committee agreed that they can help to reduce the risk of PE when therapeutic anticoagulation cannot be given. The committee also had concerns about the inherent risks involved in using IVC filters, including the invasive nature of the procedure for placing them and the potential for complications such as migration or fracture of the filter. In light of this, the committee agreed that the evidence was not sufficient to retain the 2012 recommendation to offer temporary IVC filters to people who cannot have anticoagulation treatment. However, they recognised that these people have a high risk of recurrent VTE and a limited number of alternative treatments, so agreed that a recommendation to consider an IVC filter for them is justified. They also agreed to retain the 2012 advice to remove the IVC filter when anticoagulation is no longer contraindicated, adding that anticoagulation treatment should be established before the IVC filter is removed.
There was very limited evidence on the use of IVC filters for people who have a PE while taking anticoagulation treatment for an initial proximal DVT or PE. The evidence suggested a reduction in short-term mortality from all causes in people in this group who had an IVC filter fitted. Because of the limited amount of evidence, the committee agreed that IVC filters could be considered for people in this group only after problems with adherence or other causes of hypercoagulability have been excluded, and different anticoagulation treatments or treatment regimens have been explored. The committee reasoned that, in many cases, optimising anticoagulation treatment will obviate the need for an IVC filter.
The committee noted that the type of IVC filter used is the same whether it is intended to be temporary or permanent. They agreed that, although the decision between a temporary and permanent IVC filter would be made on a case-by-case basis, most IVC filters are likely to be temporary. It is therefore prudent to have a plan in place for removal of the IVC filter at the time of fitting it, to ensure it is removed promptly when no longer needed.
It is expected that the overall impact of the recommendations will be to reduce the use of IVC filters. For people with VTE at acute risk of thrombosis, clinicians may fit an IVC filter as part of a clinical trial.
Unprovoked VTE is associated with an increased risk of cancer, which may be undiagnosed when the VTE occurs. The committee agreed that a physical examination and review of medical history (including previous investigations such as imaging) are worthwhile precautions for people who have had an apparently unprovoked DVT or PE. However, the evidence did not show any benefit from further investigations for cancer for people who have no signs or symptoms. Moreover, these investigations can be costly, time consuming, potentially invasive or pose a radiation risk, and cause anxiety. The committee therefore agreed that further investigations for cancer should not be offered to people without relevant signs or symptoms.
A physical examination and a review of medical history is current practice for people with unprovoked DVT or PE. The recommendations can be expected to reduce costs by reducing further investigations for cancer in people without symptoms or signs.