This article is more than 5 years old. Information may no longer be current.
Blood clots, bleeding and cancer
Click Here to Manage Email Alerts
The association between cancer and thrombosis has been known for more than a century.
In 1823, Jean-Baptiste Bouillaud was one of the first physicians to describe the connection. It was ironic that Armand Trousseau, who further identified this linkage in 1865, succumbed to the problem 2 years later.
The mechanisms of the association are complex and multifactorial, but known cancer-specific factors include tumor type (eg, mucin-producing adenocarcinomas, such as those arising from lung, ovary or gastrointestinal tract, and hematological malignancies), mass effect with tumor compression/invasion of the vasculature, anticancer treatment (eg, estrogens, radiotherapy), and associated chronic inflammation.
The combination of hormones (eg, tamoxifen) plus chemotherapy further increases risk for thrombosis. These associations may be compounded by more generic issues, such as a family history of thrombosis or genetic factors like factor V Leiden mutation, past history, use of other hormonal therapies (oral contraceptives), sedentary lifestyle, obesity and smoking.
Estimating the risk
It has been estimated that 20% to 30% of initial thromboses are cancer-associated. However, the converse does not seem exactly to apply when investigating patients with thrombosis for the occult presence of cancer — only about 4% to 5% of such patients turn out to have cancer after immediate investigation. Long-term follow-up or surveillance yields between 17% to 23% of patients who eventually develop cancer.
It has been estimated that patients with cancer have a four- to sixfold increased risk for thrombosis compared with the general community.
The management of established thrombosis in patients with cancer has been described extensively and does not merit a detailed description here.
Much more vexing has been the issue of whether to use antithrombotic prophylaxis for patients with cancer.
There has been an extensive and confusing literature with discordant results for decades. Randomized trials of low-molecular-weight heparin vs. placebo showed a reduction in the incidence of blood clots, but the absolute risk was seen as too low to lead to recommendation for routine prophylaxis, especially in view of the costs and inconvenience associated with daily injections. Cochrane provided a useful meta-analysis of nine major trials, concluding that thromboprophylaxis significantly reduced the incidence of symptomatic venous thrombosis but is associated with an increase in bleeding events. It concluded that such treatment should be directed to patients with cancer at high risk for developing blood clots.
Several sets of data have been produced using single-parameter biomarkers, but in recent times more complex risk-prediction models have gained favor. Khorana and colleagues have developed an algorithm based on high platelet count, high hemoglobin, use of red cell growth factors, high leukocyte count and high BMI, producing low- and high-risk groups, with respective thrombosis rates of less than 1% vs. around 7%. This approach has been widely adopted, albeit with modification.
AVERT and CASSINI
Two important randomized trials were recently reported in The New England Journal of Medicine that showed improved outcomes with thromboprophylaxis, results that may be practice-changing.
In the AVERT trial, Carrier and colleagues randomly assigned 574 patients to treatment with 2.5 mg twice-daily apixaban (Eliquis; Bristol-Myers Squibb, Pfizer) — a potent, orally administered factor Xa inhibitor — or placebo. Venous clots occurred in 12 of 288 patients treated with apixaban vs. 28 of 275 patients in the placebo group (HR = 0.41; P < .001). However, significant hemorrhage occurred respectively in 10 (3.5%) and five (1.8%) cases.
There was no survival benefit, although analysis of this endpoint would have been confounded by the presence of advanced cancer in patients in both arms, and perhaps the absolute numbers of events in each arm.
Khorana and colleagues reported results from a parallel study, the CASSINI trial (see related article).
This double-blind, randomized trial focused on intermediate/high-risk ambulatory patients with cancer (Khorana score of > 2) and involved the use of a daily dose of 10 mg rivaroxaban (Xarelto, Janssen) vs. placebo. The composite primary endpoint consisted of objectively confirmed proximal deep vein thrombosis in a lower limb, pulmonary embolism, symptomatic DVT in an upper limb or distal DVT in a lower limb, or death of venous thromboembolism in patients followed up to 180 days; however, researchers assessed the same endpoint during the dosing period as a prespecified supportive analysis. Significant hemorrhage served as the primary safety endpoint.
Among 841 randomly assigned patients, the primary endpoint occurred in 25 (6%) of 420 patients receiving rivaroxaban vs. 37 (8.8%) of 421 patients in the placebo group (HR = 0.66; 95% CI, 0.4-1.09). The on-treatment analysis showed the primary endpoint in 11 vs. 27 patients (HR = 0.4; 95% CI, 0.2-0.8). Major bleeding occurred in 2% of the rivaroxaban group and 1% of the placebo group (HR = 1.96; 95% CI, 0.59-6.49).
Drawing conclusions
I imagine that our friends in the Cochrane group will find another publication by meta-analyzing these two studies but, in the interim, what can we conclude?
It is clear that there is a reduction of cancer-associated clot formation, but is it worthwhile?
The two trials have similar results, but the use of the composite endpoint in the CASSINI trial will certainly have obfuscated interpretation to some extent. There was significant discontinuation of medication, both in the rivaroxaban and placebo groups in CASSINI, which may fit with the population of patients with advanced cancer.
Ultimately, the decision to use this type of prophylaxis comes down to interpretation of risk in each trial vs. the confounding variables. Ideally, each trial might have been designed with greater power, and probably would have resolved some of the questions that have emerged.
My own take is that this issue of NEJM gives pretty strong support to the use of one of these agents prophylactically for patients with truly high risk, and the situation for those with intermediate risk for clot formation is less clear.
The move from parenteral therapy to oral therapy removes one significant problem. The paucity of deaths from hemorrhage is important, but we should not forget that these patients were not really in the “real-world” setting, as they were the subjects of active investigation by teams of committed investigators with more than average clinical support for monitoring of side effects and other problems.
References:
Carrier M, et al. N Engl J Med. 2019;doi:10.1056/NEJMoa1814468.
Khorana AA, et al. Blood. 2008;doi:10.1182/blood-2007-10-116327.
Khorana AA, et al. N Engl J Med. 2019;doi:10.1056/NEJMoa1814630.
For more information:
Derek Raghavan, MD, PhD, FACP, FRACP, FASCO, is HemOnc Today’s Chief Medical Editor for Oncology. He also is president of Levine Cancer Institute at Atrium Health. He can be reached at derek.raghavan@atriumhealth.org.
Disclosure: Raghavan reports no relevant financial disclosures.