Subscribe to eTOC

Monitoring for Atrial Fibrillation After Cryptogenic Stroke
Which Patients, and for How Long?

ARTICLE IN BRIEF

Figure

MODEL PREDICTIONS of AF detection within 12 months (A) and 35 months (B) of insertable cardiac monitoring by age, PR interval, and use of medications that can prolong the PR interval.

In a new analysis, investigators reported that increased age and prolonged PR interval at baseline were the best predictors of the risk of atrial fibrillation after cryptogenic stroke.

After a cryptogenic stroke, which patients should receive extended monitoring for atrial fibrillation (AF)? That question assumed more prominence after the 2014 publication of two major trials, CRYSTAL AF and EMBRACE, both of which showed that long-term monitoring could identify many instances of AF in patients who left the hospital without any evidence of faulty heart rhythm.

Now, an analysis of CRYSTAL AF patients who received long-term monitoring reports that increased age and prolonged PR interval at baseline are the best predictors of risk of AF after cryptogenic stroke. But the effect is modest, and experts are divided over the practical consequence of the results. The study was published in the December 18 online edition of Neurology.

Figure

DR. VINCENT THIJS: “If you cannot implant everyone with a device after cryptogenic stroke, then the question becomes whether you want to pick up as much atrial fibrillation as possible, in which case one would target the elderly and those with a prolonged PR interval, or whether it would be better to target younger patients, who have more years of life and quality of life to gain if stroke can be averted. It becomes an ethical question.”

The device used in the CRYSTAL AF trial was an insertable cardiac monitor whose battery life allows it to record for up to three years, with data recovered remotely. The advantage of the device, according to the lead author of the new study, Vincent Thijs, MD, PhD, a neurologist at Austin Health in Heidelberg, Victoria, Australia, is that AF not detected in the clinic may be asymptomatic and intermittent, “so you need to be testing exactly at the moment it occurs.”

STUDY PROTOCOLS, FINDINGS

In the trial, patients with cryptogenic stroke were randomized to standard follow-up or to monitoring with the implantable device and were seen at months one, six, and 12, then every 12 months thereafter. Dr. Thijs analyzed baseline characteristics of the 221 patients who received the device. The participants had a mean age of 61 years, 88 percent were white, and 64 percent were male.

Atrial fibrillation occurred in 42 patients: 29 during the first 12 months and 13 between months 12 and 36. Using a multivariate model, Dr. Thijs and colleagues found that older age and longer PR interval were independently associated with a greater risk of stroke, while diabetes and premature atrial contractions were not.

Each 10-year increase in age was associated with a hazard ratio of 1.9 (95% CI, 1.3-2.8; p<0.0009), and a prolonged PR interval was associated with a hazard ratio of 1.3 (95% CI, 1.2-1.4; p<0.0001).

While the associations were significant, Dr. Thijs said, they had “only moderately predictive capacity” in the study, and individual prediction of AF “remains elusive.” However, he said, “the main message is that we should be aware of atrial fibrillation in cryptogenic stroke, and we should try all means to identify it. And the longer you look, the more you will find.”

Dr. Thijs said there are two options to determine who should be monitored. “If you cannot implant everyone with a device after cryptogenic stroke, then the question becomes whether you want to pick up as much atrial fibrillation as possible, in which case one would target the elderly and those with a prolonged PR interval, or whether it would be better to target younger patients, who have more years of life and quality of life to gain if stroke can be averted. It becomes an ethical question.”

Underlying that question is one of allocation of health care dollars, both for the device and for the extra labor needed to monitor it. Data from the United Kingdom presented in 2015 at the International Stroke Conference but not yet published suggest that the implantable device is “not cost-saving, but is cost-effective,” Dr. Thijs said, when the extra years of quality of life gained by avoiding a stroke are taken into account.

But even more fundamentally, the need for monitoring and the question of who should be monitored arise from the treatment choices that follow detection of AF in post-stroke patients, namely the decision to use an anticoagulant rather than an antiplatelet agent. The risk of intracranial bleeding currently limits the use of anticoagulants to the patients most at risk of a second stroke from AF.

Figure

PREDICTED VS. OBSERVED AF detection rates are shown for six subgroups of insertable cardiac monitor arm patients based on Cox model prediction of AF detection rates. The predicted AF detection rate at the end of the follow-up period based on the mean Cox linear predictor for each subgroup is plotted (x-axis) against the observed rate determined by the Kaplan-Meier estimate (y-axis). The predicted rate for the minimum and maximum Cox linear predictor among patients in the subgroup are also plotted to describe the range of predictions within the subgroups. The predicted vs. observed calibration curves are shown for AF dection within 12 months (A) and within 36 months (B).

EXPERTS COMMENT

“I think for neurologists the most important point is that if you are going to monitor your patients, particularly for a long time, you have to be committed to using the data for determining treatment,” said Alejandro Rabinstein, MD, FAAN, a professor of neurology at the Mayo Clinic in Rochester, MN, who was not involved with the study. “You should monitor patients in whom you have not established another mechanism after an extensive search and in whom you think that anticoagulation would be a good idea if you discover atrial fibrillation. If you are not going to use an anticoagulant, it doesn't matter whether or not they have atrial fibrillation.”

For Wade S. Smith, MD, PhD, a professor of neurology at the University of California, San Francisco, the study falls short of being a definitive tool for decision-making, as the authors themselves emphasized. “If we had a good predictive tool to know who was most at risk, maybe we could put those patients on anticoagulants while we performed long-term monitoring,” he said. “But while the current study is commendable, it doesn't tip the balance enough to change practice, in my opinion.”

“I think the data are pretty clear that the longer you look, the more you are going to find,” said Seemant Chaturvedi, MD, a professor of clinical neurology at the University of Miami School of Medicine in Florida. “But the question is how to make efficient use of resources, since we don't want prolonged monitoring on everybody, especially if it will be low-yield. In a large number of patients, if you monitor for four weeks and you don't find any atrial fibrillation, that is going to be sufficient, but there may be some special circumstances where prolonged monitoring is desirable, and I think this paper provides a couple of clues.”

Age as a risk factor was not surprising, he said, but the prolonged PR interval was “something neurologists don't give a lot of thought to, and that may be a nice pearl the neurologist can include in their decision-making.”

“This is an important and well-done study, which is in agreement with prior studies that [show] occult paroxysmal atrial fibrillation is ongoing in some patients,” said Steven R. Messe, MD, FAAN, an associate professor of neurology at the Hospital of the University of Pennsylvania in Philadelphia. “Nonetheless, while the device is unobtrusive and generally well tolerated, there is some cost associated, and it is not yet clear it is the best way to use health care resources. In addition, there is still an unanswered question of how relevant small episodes of atrial fibrillation are when they are identified. Are they always the cause of the stroke? Do they always require anticoagulants? Oftentimes they probably do, but we don't yet know for sure.”

Studies to answer those questions are ongoing. In addition, studies of the novel anticoagulants dabigatran, rivaroxaban, and apixaban in patients with embolic stroke of unknown source — a term that may replace “cryptogenic stroke” in the near future, experts told Neurology Today — may render the question of long-term monitoring moot, if they can be shown to be safe enough to prescribe to most patients, including those without evidence of atrial fibrillation. But those studies are just getting underway, with results not expected for several years, leaving neurologists with the important and still largely unanswered question of which patients need monitoring, and for how long, after a cryptogenic stroke.

EXPERTS: ON WHO TO MONITOR FOR ATRIAL FIBRILLATION

Figure

DR. ALEJANDRO RABINSTEIN: “You should monitor patients in whom you have not established another mechanism after an extensive search and in whom you think that anticoagulation would be a good idea if you discover atrial fibrillation. If you are not going to use an anticoagulant, it doesnt matter whether or not they have atrial fibrillation.”

Figure

DR. SEEMANT CHATURVEDI: “I think the data are pretty clear that the longer you look, the more you are going to find. But the question is how to make efficient use of resources, since we dont want prolonged monitoring on everybody, especially if it will be low-yield. In a large number of patients, if you monitor for four weeks and you dont find any atrial fibrillation, that is going to be sufficient, but there may be some special circumstances where prolonged monitoring is desirable, and I think this paper provides a couple of clues.”

Figure

DR. WADE S. SMITH said the study falls short of being a definitive tool for decision-making, as the authors themselves emphasized. He added: “If we had a good predictive tool to know who was most at risk, maybe we could put those patients on anticoagulants while we performed long-term monitoring. But while the current study is commendable, it doesnt tip the balance enough to change practice, in my opinion.”

Figure

DR. STEVEN R. MESSE: “This is an important and well-done study, which is in agreement with prior studies that [show] occult paroxysmal atrial fibrillation is ongoing in some patients. Nonetheless, while the device is unobtrusive and generally well tolerated, there is some cost associated, and it is not yet clear it is the best way to use health care resources. In addition, there is still an unanswered question of how relevant small episodes of atrial fibrillation are when they are identified. Are they always the cause of the stroke? Do they always require anticoagulants? Oftentimes they probably do, but we dont yet know for sure.”

LINK UP FOR MORE INFORMATION:

•. Thijs V, Brachmann J, Morillo CA, et al. Predictors for atrial fibrillation detection after cryptogenic stroke: Results from CRYSTAL AF http://www.neurology.org/content/early/2015/12/18/WNL.0000000000002282.abstract. Neurology 2015; Epub 2015 Dec. 18.
    •. Diener ST, Passman RS, Di Lazzaro V, et al. for the CRYSTAL AF Investigatorrs. Cryptogenic stroke and underlying atrial fibrillation http://www.nejm.org/doi/full/10.1056/NEJMoa1313600. N Engl Med 2014; 370(26): 2478–2486.
      •. Gladstone DJ, Spring M, Dorian P, et al. for the EMBRACE Investigators and Coordinators. Atrial fibrillation in patients with cryptogenic stroke http://www.nejm.org/doi/full/10.1056/NEJMoa1311376. N Engl J Med 2014; 370: 2467–2477.