Long-term Anticoagulation After Unprovoked VTE: New Insights

To guide treatment choices, a review offers some hard numbers on the likelihood of later bleeds or subsequent clots.

Long-term Anticoagulation After Unprovoked VTE: New Insights

The hazards associated with keeping patients on long-term anticoagulation after an unprovoked venous thromboembolism (VTE) are “considerable,” but the authors of a new study say they have some hard numbers to help clinicians and patients understand the relative risks and benefits of continuing therapy.

Current guidelines suggest that a wide range of major risk factors for bleeding be considered before deciding to keep a patient on either a vitamin K antagonist (VKA) or direct oral anticoagulant (DOAC) beyond the first 3 to 6 months. But as Faizan Khan, MSc (University of Ottawa, Canada), and colleagues point out, the true risks of bleeding for people whose VTE can’t be traced back to common triggers such as surgery or cancer are unclear.

“About half of patients who develop deep vein thrombosis or pulmonary embolism develop it without a strong provoking risk factor, and in those patients the current recommendations are to consider lifelong or indefinite anticoagulants after a first clot,” so long as they are not at high risk for a bleed, senior author Marc A. Rodger, MD (McGill University, Montreal, Canada), told TCTMD. However, trying to convince particularly younger patients to take a lifelong medication, without solid data on whether the long-term benefits outweigh the risks, can make for a tough discussion, he said.

The aim of their review, Rodger said, was to try to get some precise numbers around the true long-term likelihood of bleeding for those patients, while also identifying which factors are most important in determining that risk.

“We've got lots of data that tell us about the sort of short-term risks of bleeding during the initial treatment of the clot, and even for the first year or two,” he noted. “But when we're making these lifelong decisions for patients, we didn't think that we could or should be extrapolating early bleeding-risk data to the long term, because we know that the bleeding risk goes down in anticoagulant-experienced patients with time.”

Risk Beyond 1 Year

In the review, which was published last week in the Annals of Internal Medicine, Khan et al combined data from 14 randomized controlled trials and 13 cohort studies for a total of 9,982 VTE patients treated with a VKA and 7,220 treated with a DOAC. Major bleeds occurred at a rate of 1.74 per 100 person-years in patients treated with VKAs and 1.12  in those treated with DOACs.

By 5 years, the cumulative incidence of major bleeding was 6.3% with VKAs; however, the data were “insufficient” to assess the risk with DOACs beyond 1 year. When treatment groups were combined, however, investigators reported significantly higher rates of major bleeding among patients older than 65 years, creatinine clearance less than 50 mL/min, prior bleeding, concomitant use of an antiplatelet, or a hemoglobin level less than 100 g/L. Importantly, case fatality rates associated with major bleeding were high, ranging from 8.3% in patients taking VKAs to 9.7% in patients taking a DOAC.

Current guidelines offer what Rodger described as an exhaustive table of bleeding risk factors, several pages long, that’s not practical for decision-making. “That is a huge criticism from practicing physicians of the ‘laundry-list approach.’ . . . You can't ask me to remember that in practice,” he said. “I'm not going to go back to the paper every time to try and risk-categorize people.”

The laundry list also is not useful for evaluating the weight of a given risk factor in an individual patient, he added. Instead, the authors suggest, the presence of one or more of the five risk factors highlighted in their analysis can help identify patients for whom the risks of extending anticoagulation beyond the initial 3 to 6 months may outweigh the benefits, and they put some hard numbers to that uncertainty.

Armed with these figures, he explained, “you can guide the patient with data that tells them that their risk of developing a new clot is 40% to 50% in the next 10 years if we stopped your blood thinners, and 1% or 2% if we continue in your blood thinners. But the counterbalancing [thing] is that your risk of bleeding in the next 10 years is going to be somewhere in the 5% to 10% range. So let's look at your risk factors and think about whether you're going to be even higher risk than that sort of average, 5% or 10%, and see if that should influence whether we should keep you on forever or not.”

Khan, also speaking with TCTMD, pointed out that the trials included in the analysis date back 10 to 15 years. Combining them allowed for what he called the “best available estimates” for physicians to use in discussing risks and benefits with patients, as well as spotlighting risk factors of greatest concern.

But he stressed that despite the fact that DOACs have been used in clinical practice for extended treatment, and are approved for this indication, their safety beyond 1 year has not been well studied. “So that’s one piece of information that we would have liked to have had available in our meta-analysis, and something that should also be the focus of future research,” Khan said.

A Welcome Contribution

Commenting on the analysis for TCTMD, Geoffrey Barnes, MD (University of Michigan, Ann Arbor), a member of the American College of Cardiology’s Peripheral Vascular Disease Section, zeroed in on the dearth of long-term DOAC bleeding data, noting that the paper importantly draws attention to this gap.

“The other thing it does is: it helps us identify a couple of key groups for whom the bleeding risk seems to be notably elevated,” Barnes said. “That can help us as clinicians, as we start to think in whom might that bleeding risk exceed the thrombotic risk. . . . Maybe these are people in whom we would think of not giving lifelong anticoagulation, or not continuing anticoagulation therapy.”

Like Rodger, Barnes said that the comprehensive list of risk factors offered by the guidelines is not useful in practice. “As a clinician, you look at those lists that are 20 items long—nobody is going to remember that. They’re very hard to use clinically.” What this paper makes clear, he said, is that “there are just a couple key elements that we can all keep in mind.”

And while some of these risk factors, such as age, can’t be altered, others may allow for interventions that could lower bleeding risk, thereby changing the equation. Moreover, said Barnes, the paper makes the point that bleeding risk changes with time. “That tells us that we have to continually reassess bleeding risk to decide if that therapy is appropriate,” he added.

Rodger, to TCTMD, explained that the current paper is part of a larger project that he and Khan are doing which seeks to pool data from all available sources to create a decision analysis for large groups of patients as well as subgroups of patients clarifying the safety and efficacy of remaining on long-term therapy. “We're hoping that that will be ultimately guideline-relevant,” Rodger said, “because the guidelines currently are honestly more expert opinion and not a formal synthesis of data and not a formal decision analysis.”

Shelley Wood is Managing Editor of TCTMD and the Editorial Director at CRF. She did her undergraduate degree at McGill…

Read Full Bio
  • Khan F, Tritschler T, Kimpton M, et al. Long-term risk for major bleeding during extended oral anticoagulant therapy for first unprovoked venous thromboembolism. Annals Intern Med. 2021;Epub ahead of print.

  • Khan and Rodger report being members of the CanVECTOR Network which receives grant funding from the Canadian Institutes of Health Research.