Anticoagulation Suboptimal in Patients With Cancer and AFib

Patrice Wendling

October 09, 2018

A large database study suggests the likelihood of seeing a cardiologist after an atrial fibrillation (AF) diagnosis is consistently lower among patients with any active cancer or history of malignancy than in those without cancer.

Early referral to cardiology specialists in patients with cancer, however, was associated with an increase in prescriptions for oral anticoagulants and an 11% lower relative risk for stroke (RR, 0.89; 95% CI, 0.81 - 0.99) without an uptick in bleeding risk.

"An 11% reduced risk of stroke is clearly not a negligible risk," lead author, Wesley T. O'Neal, MD, MPH, Emory University School of Medicine, Atlanta, told | Medscape Cardiology. "We thought this was very important given that patients with cancer and atrial fibrillation have a higher risk of stroke clearly than those who don't develop atrial fibrillation."

The study was published online October 8 in the Journal of the American College of Cardiology.

"The big overall take-home message is that there are gaps in care for patients with cancer and atrial fibrillation and that more can be done to optimize the care of these patients," Sean T. Chen, MD, Duke University Medical Center, Durham, North Carolina, who wrote an accompanying editorial, said in an interview.

He emphasized that patients with cancer represent a very heterogeneous population and that even within this study, there was significant variation in rates of prescribing among different cancer subtypes.

"The other thing I found interesting was despite higher rates of anticoagulation use among patients who are seen by cardiologists, there wasn't really a higher bleeding risk," Chen said. "This doesn't mean necessarily that every patient who has cancer and gets diagnosed with atrial fibrillation should be put on anticoagulation, but it raises the point that more can be done."

Similarly, Ana Barac, MD, MedStar Heart and Vascular Institute and Georgetown University, Washington, DC, and chair of the American College of Cardiology's cardio-oncology section, told | Medscape Cardiology, "It's an important red flag but we should not rush into interpreting it as, 'Oh, we should now anticoagulate all the patients or we should not.' The truth is we don't know."

She also stressed that the population is heterogeneous and the literature very sparse. There are data to suggest the risk for AF is greater, for example, in lung cancer or after bone marrow transplant than in breast cancer, but the risk of anticoagulation also varies. Although most patients with cancer are hypercoagulable, Barac noted that the risk of bleeding may be significantly elevated, for example, in patients undergoing procedures, such as diagnostic biopsies, and in those at risk of bleeding from the cancer site, which is a "huge concern for all patients with gastrointestinal tract cancers and all cancers with potential metastases to the brain."

Although a recent randomized trial showed that the novel oral anticoagulant edoxaban (Savaysa, Daiichi Sankyo) was noninferior to injections of dalteparin (Fragmin, Pfizer) for cancer-related venous thromboembolism, Barac was unaware of any randomized controlled trials that have specifically looked at AF in patients with cancer, much less that evaluated the newer direct oral anticoagulants (DOACs).

Barac and Chen also point out that the CHA₂DS₂-VASc and HAS-BLED scores — routinely used to assess stroke risk and bleeding risk, respectively, in patients with AF — have never been validated in cancer patients.

"What this paper points to is that we have a chance to improve outcomes, but only if we have an interdisciplinary discussion," Barac said. "I don't think it is a kind of recipe to anticoagulate — that is perhaps my biggest fear."

"This is truly an observational study that uses ICD-9 codes, with all the dangers there," she said. "It provides a first step towards practice change that needs to start with collaboration in the cardio-oncology clinic and investigation of cancer-specific bleeding and thrombosis risks in the setting of atrial fibrillation."

MarketScan Data

Using the Truven Health MarketScan databases from January 2009 through December 2014, O'Neal and his colleagues identified 388,045 patients with an ICD-9 code for nonvalvular AF, of which 64,016 (17%) had a history of cancer. Analyses were stratified by cancer status (remote or active treatment in the 6 months prior to AF diagnosis) and cancer type (prostate, 26%; breast, 19%; lung, 14%; hematologic, 14%; colon, 12%; and pancreatic, 2%).

Patients with cancer were older than those without cancer (74 vs 67 years; < .001) and significantly more likely to have heart failure, hypertension, diabetes, stroke, MI, and bleeding, and to have higher CHA₂DS₂-VASc (3.7 vs 3.0) and HAS-BLED (2.4 vs 1.8) scores. Patients with cancer were also more likely to fill prescriptions for cardiovascular medications.

Patients who had a history of cancer were less likely to see a cardiology provider after their AF diagnosis than those without cancer (54% vs 62%; RR, 0.92; 95% CI, 0.91 - 0.93). This trend was stronger among patients with active vs remote cancer, although no differences in cardiology care were seen among patients with breast or prostate cancer, which are perceived to be less aggressive forms of cancer, O'Neal said.

Anticoagulation prescription fills also were lower in patients with cancer than in those without cancer (31% vs 35%; RR, 0.89; 95% CI, 0.88 - 0.90). Prescription fill rates varied from a high of 35% in prostate cancer to a low of 18% in pancreatic cancer, but overall results were similar for all cancer subtypes except breast cancer, based on active vs remote cancer status; and in a propensity-score matched cohort.

Among patients with cancer, those seen by a cardiology provider were more likely to fill prescriptions for anticoagulants than those not seen by a cardiologist (RR, 1.48; 95% CI, 1.45 – 1.52). The results did not vary between any of the cancers, by active vs remote cancer status, or by HAS-BLED score (≥3 vs <3).

During a mean follow-up of 1.1 years after AF diagnosis, the stroke incidence rate was 17.3 per 1000 person-years for patients seem by primary care physicians and 15.0 per 1000 person-years for those seen by cardiologists.

The major bleeding incidence rates for those seen by a primary care provider vs a cardiologist were 20.5 vs 20.7 per 1000 person-years, respectively (RR, 1.04; 95% CI, 0.95 -1.13).

Additionally, AF patients with cancer seen by a cardiologist were more likely to be hospitalized for heart failure and AF, than those seen by primary care providers.

"What could be possibly happening here, and this is largely speculative, is that patients are likely perceived to have poor survival, therefore their cardiology provider or primary care provider may not treat their atrial fibrillation as aggressively as if they had never had a history of cancer," O'Neal said.

"I think some of the data can speak to this because if patients had more aggressive forms of cancer, they clearly were less likely to see a cardiology provider than those with less aggressive forms of cancer; and that mirrored what we were able to see if they filled a prescription for a DOAC or not."

In the Millions and Climbing

O'Neal and colleagues point out that because of advances in cancer detection and treatment, the number of Americans surviving cancer will rise to nearly 19 million over the next 5 years.

"We're going to see more of these patients and we're going to have to develop targeted treatment strategies to reduce their risk of cardiovascular complications, especially atrial fibrillation, which is very common in cancer patients," O'Neal said.

"It's overwhelming to be diagnosed with any cancer from an oncology standpoint and then to layer in the possibility of future cardiovascular complications, I don't think that's on the table."

One of the challenges in this population is that these patients don't live in the cardiovascular space, Barac said. "These are oncology patients, so there is much less interest in atrial fibrillation but also probably even when they are seen, much more awareness of the potential risks. These patients are more likely to bleed."

"That said, in a number of malignancies, once atrial fibrillation is confirmed at least in clinical practice, oncologists actually tend to be very agreeable to anticoagulate because these patients are probably at higher risk of clotting."

"Whatever it's called — cardio-oncology or onco-cardiology — it would be helpful to train more cardiologists to be more comfortable in the cancer field, the same way that oncologists need to be aware that new treatments, such as DOACs, are available for AF and that AF can put the patient at risk for heart failure and jeopardize outcomes," Barac said.

It would be ideal if a cardio-oncologist was available, but in a lot of different community settings that is not the case, Chen observed.

"Referral to a general cardiologist can still be helpful in having this conversation in terms of overall risk assessment and whether a patient would benefit from anticoagulation with warfarin or DOACs," he said. "Different providers will have different comfort levels in prescribing these medications, so I think it's an ongoing conversation that patients will need to have with their primary care provider, their local cardiologist, and oncologist."

In the editorial, Chen and his coauthor, Chiara Melloni, MD, MHS, Duke University Research Institute, note that the study was not powered to detect differences in provider outcomes among specific cancers and that the mechanisms by which cardiology involvement leads to lower stroke rates likely extend well beyond rates of anticoagulation use.

Use of rhythm control strategies with catheter ablation or direct cardioversion and optimization of cardiovascular comorbidities also play a role. Future studies also would benefit from examining potential differences in anticoagulation adherence between patients seen by different specialties, they note.

The research was supported by awards from the National Heart, Lung, and Blood Institute, National Institute on Aging, and America Heart Association. The authors report having no relevant financial relationships. Chen reports no having no relevant financial relationships. Melloni has received research grants from Amgen, AstraZeneca, Bristol-Myers Squibb, Ferring Pharmaceuticals, GlaxoSmithKline, Luitpold Pharmaceuticals, Merck, Roche Group, Sanofi, St. Jude Medical, and the NICHD. Barac reports having no relevant conflicts of interest.

J Am Coll Cardiol. 2018;72:1913-1922 and 1923-1925. Abstract, Editorial

Follow Patrice Wendling on Twitter: @pwendl. For more from, follow us on Twitter and Facebook.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: