FAQs From Oncologists

Gerald A. Soff, MD
Chief of the hematology service, Memorial Sloan Kettering Cancer Center, New York.

In this first edition of HemOnc Link, Gerald Soff, MD, sheds light on some of the most common questions that hematologists are asked by their solid tumor oncologist colleagues.

Cancer-Associated Thrombosis: When and How To Treat

Cancer-associated thrombosis is common, occurring in approximately 15% of patients with cancer, with rates varying depending on the cancer type. The consensus is that virtually all patients should be anticoagulated, so the first question we receive as hematologists is often about the choice of anticoagulant – a direct oral anticoagulant (DOAC) or low-molecular-weight heparin (LMWH).

Twenty-five years ago, the anticoagulation question was much simpler: Warfarin or no warfarin? Then, the LMWHs came along, and there was a battle between the science showing LMWH was more effective and the question of whether patients would adhere better to an oral drug. The choice of drug now requires a lot of expertise and is a nuanced decision.

In general, the DOACs are all similar, but there are a few key differences. For example, edoxaban has a different spectrum of drug interactions from the others, while apixaban seems to have the least renal clearance so might be an option for some patients with renal insufficiency. All DOACs have very good safety and efficacy data, and, of course, they have the advantage of oral administration, often with once-daily dosing. However, there are no randomized data comparing the safety and efficacy of any of the DOACs against each other – particularly in patients with cancer.

The literature has demonstrated that patients who receive DOACs are at an increased risk of urinary tract and gastrointestinal bleeding. Based on the data and our experiences at our center, I would recommend that, as long as the patient does not have a urinary tract infection or digestive tract abnormality, he or she should be able to receive a DOAC.

I would recommend LMWH for any patient with renal insufficiency, as LMWHs have clear dosing guidelines for these patients. The guidance for DOACs is less complete and has not been validated as thoroughly. However, the FDA provided guidance for the use of apixaban in patients with end-stage kidney disease on hemodialysis, based on limited pharmacokinetic data, so that may be an appropriate choice.

We also must consider drug interactions. Certain cancer treatments and antifungal drugs inhibit the CYP3A4 enzyme and have strong interactions with DOACs, such as rivaroxaban and apixaban. In those cases, we would recommend edoxaban or dabigatran.

When Can I Stop Anticoagulation?

For a patient with active cancer or who is continuing to receive cancer-directed therapy, we recommend ongoing anticoagulation.

However, patients may be able to discontinue anticoagulation at some point. Take, for example, a woman with ovarian cancer who has undergone surgery and develops a DVT for which she receives anticoagulation. Following surgery, she begins chemotherapy and is closely monitored. After several cycles, the cancer is undetectable and she discontinues chemotherapy.

At this point, essentially, the parameters that put her at risk for the thrombosis are no longer there: She doesn’t have measurable active cancer and she is not receiving chemotherapy. So, as long as the DVT has been treated for a sufficient length of time (typically at least 6 months), we would recommend stopping the anticoagulation.

Now, there may still be undetectable microscopic disease that might lead to a relapse later on, but if the seromarkers and the imaging studies are negative, we call that “no evidence of disease,” or NED. The patient may not be cured long-term, but, as long as he or she is NED and off cancer-directed therapy, we would discontinue anticoagulation.

It is less clear what to do for patients who are on long-term maintenance therapy, such as hormonal therapy or a poly-ADP ribose polymerase (PARP) inhibitor, and this needs to be individualized.

Does a History of Clot Affect Prophylactic Anticoagulation Decisions?

In the scenario of a patient diagnosed with cancer who has experienced a DVT or PE within approximately 6 months before the diagnosis, I would argue that there was probably cancer present but not detected. In that case, the blood clot would be considered cancer related.

However, let’s assume a scenario in which a 60-year-old woman with metastatic breast cancer had a DVT when she was 25 and on oral contraceptives. That DVT is certainly not related to her cancer.

What we don’t know yet is what to do with that information. So, if a woman has active breast cancer and had a thrombosis in the distant past, I would not anticoagulate her prophylactically based on that remote history alone. However, if she had a DVT in the previous 6 months and is receiving anticoagulation, then develops the cancer, we would choose to continue the anticoagulation.

Is Anticoagulation Safe in a Patient With Brain Tumors?

Jeffrey Zwicker, MD, has published several articles evaluating anticoagulation and the risks of intracranial hemorrhage (ICH) in the setting of patients with brain metastases. He and his colleagues have published retrospective data, which are the best data available to us, for now.

In a 2015 study, Dr. Zwicker and colleagues assessed the risk for ICH among patients with cancer who have brain metastases. The ICH risk was 4-fold higher in patients with melanoma or renal cell carcinoma than those with lung cancer, but the risk was not influenced by the use of the LMWH enoxaparin – which was the only anticoagulant available for patients with cancer at the time. The evidence appeared to support the safety of therapeutic anticoagulation in this population.1

The consensus is that virtually all patients [with cancer-associated thrombosis] should be anticoagulated, so the first question hematologists receive is often about the choice of anticoagulant – a DOAC or LMWH.

A second paper, published in 2017, demonstrated a different pattern: In patients with primary brain tumors (not metastatic to the brain), treatment with enoxaparin was associated with a three times higher likelihood of experiencing a major ICH, compared with those who were not anticoagulated.2

These studies demonstrate that there is a difference between cancers metastatic to the brain and cancers arising in the brain. Some brain metastases already have a hemorrhagic component. Generally, in those cases, I would forgo anticoagulation. However, if the patient’s brain metastases have no known hemorrhagic components, then I would defer to Dr. Zwicker’s 2015 paper, which supported anticoagulation.

In a third paper, published in 2019, the team evaluated rates of ICH in patients with brain tumors who received DOACs. They found that, if a clinician decides to anticoagulate a patient with primary or metastatic disease, the DOACs seem to be at least as safe as the LMWHs.3

Based on available evidence, it seems that DOACs would be a reasonable choice once the decision has been made to anticoagulate.

Does a Venous Catheter-Related DVT Need to Be Anticoagulated?

There are three scenarios in which an upper-extremity central line-associated thrombosis is found:

  • The patient has swelling, pain, or stiffness around the neck or in the arm. Typically, there’s a symptomatic thrombosis obstructing the flow of blood.
  • The patient is asymptomatic, but the central line is no longer working well because of a clot at the end of the catheter that prevents infusion, blood withdrawal, or both.
  • A clot is identified incidentally. For instance, the patient has a CT scan that reveals a thrombus near the tip of the central line, a clot along the line, or a fibrin sheath around the catheter or in one of the veins. The patient is asymptomatic.

If the patient is symptomatic, we anticoagulate unless there is a contraindication, and usually remove the line.

If the catheter isn’t working, we would anticoagulate with an intent to save or preserve the line. If, despite anticoagulation, the line isn’t functional, then we would usually pull it. Generally, we would anticoagulate these patients for at least a month or more, although there is no consensus on how long to do so. If there’s a goal to preserve the line after treating for a symptomatic clot, we might continue anticoagulation at a prophylactic dose at least. However, there is no science to provide definitive guidance on what to do for a central line with a clot.

I might not do anything for an asymptomatic fibrin sheath, because anticoagulation has a risk of hemorrhage and the benefit of anticoagulation in this situation is unclear. I might anticoagulate if the sheath extended to the right atrium, however.

Historically, LMWH has been the drug of choice for catheter-associated thrombosis. Case series have been published showing that rivaroxaban was able to preserve most central lines. We’re comfortable using rivaroxaban in most patients with the same limitations that we do for DVT in the leg.

When Is Prophylactic Anticoagulant Needed?

The multinational CASSINI study4 and the Canadian AVERT study5 addressed the risks and benefits of prophylactic anticoagulation in cancer patients. Both studies enrolled ambulatory cancer patients who had a Khorana score6 of ≥2.

In the CASSINI study, which looked at rivaroxaban, all patients were prescreened with a Doppler ultrasound, and 4.5% of those patients had a preexisting DVT. Therefore, the numbers would not be directly comparable because 4.5% represented half of the events identified in the study.

The AVERT study, which looked at apixaban, showed a decrease in symptomatic thrombosis. CASSINI showed a reduction in all thrombosis. Both studies showed a significant reduction in thrombosis while on treatment and a small increase in the risk of major hemorrhage. Collectively, these two studies, which were similar in outcome, support the use of prophylactic anticoagulation with apixaban or rivaroxaban in most patients with Khorana scores of ≥2 who are about to initiate a new chemotherapy regimen.

If a patient has a new gastroesophageal, bladder, or kidney cancer, they’re at a high risk for hemorrhage, and regardless of the Khorana score, I would not use prophylactic anticoagulation.

Absent a contraindication, it is appropriate to consider prophylactic anticoagulation. No current clinical guidelines mandate prophylactic anticoagulation in these settings. I believe the science is moving ahead to support the use of rivaroxaban and apixaban in this setting, but it hasn’t been compelling enough to suggest it should be the new standard of care.

When Should I Consult a Hematologist?

In general, the answer depends on the oncologist’s individual level of expertise with anticoagulation and comfort with the specific scenario. It would be impossible for every oncologist to consult a hematologist about prophylactic anticoagulation each time they are about to start a patient with a marginally high Khorana score on chemotherapy, for example.

Many oncologists dual train in hematology and oncology and, if they keep up with the current scientific literature and feel comfortable making these decisions on their own, they do not need to consult a hematologist. Of course, they should also try to keep current on their local guidelines regarding the management of cancer-associated thrombosis.

On the other hand, many oncologists may not have kept up with the hematology training – or never received it.

If a patient with an uncomplicated DVT or PE does not have one of the contraindications outlined above and seems to have a fairly routine cancer-associated thrombosis, I feel that well-trained oncologists should be comfortable managing this clinical problem on their own.
If the patient has a contraindication or significant comorbidities, or if the oncologist is simply uncomfortable managing a patient, then bring in a hematologist.


  1. Donato J, Campigotto F, Uhlmann EJ, et al. Intracranial hemorrhage in patients with brain metastases treated with therapeutic enoxaparin: a matched cohort study. Blood. 2015 Jul 23;126(4):494-9.
  2. Mantia C, Uhlmann EJ, Puligandla M. Predicting the higher rate of intracranial hemorrhage in glioma patients receiving therapeutic enoxaparin. Blood. 2017 Jun 22;129(25):3379-3385.
  3. Carney BJ, Uhlmann EJ, Puligandla M. Intracranial hemorrhage with direct oral anticoagulants in patients with brain tumors. J Thromb Haemost. 2019 Jan;17(1):72-76.
    Khorana A, Soff G, Kakkar A, et al. Rivaroxaban for Thromboprophylaxis in High-Risk Ambulatory Patients with Cancer. N Engl J Med. 2019; 380:720-728.
  4. Carrier M, Abou-Nassar K, Mallick R, et al. Apixaban to Prevent Venous Thromboembolism in Patients with Cancer. N Engl J Med. 2019; 380:711-719.
  5. Khorana A, Kuderer NM, Culakova E, et al. Development and validation of a predictive model for chemotherapy-associated thrombosis. Blood. 2008; 111(10):4902-4907.