Practice Current: What is your diagnostic evaluation of cryptogenic stroke?

Current evidence

As a result of the poor understanding of the pathogenic mechanisms, and of the lack of Class I evidence to support a specific intervention over the other, the workup and treatment strategies for CS can vary considerably. In general, in case of suspected stroke, the current guidelines from the American Heart Association (AHA)/American Stroke Association (ASA) recommend at a minimum to obtain a brain CT or brain MRI, basic laboratory panels including blood counts, coagulation, electrolytes, and renal function, oxygen saturation, markers of cardiac ischemia, and an EKG.⁷ Noninvasive vascular imaging and TTE are commonly performed as well.

Several pathogenic mechanisms have been described in association with CS, and in this article we focus on 3 important mechanisms: occult atrial fibrillation (AF), occult cardiac sources, and thrombophilic states. In each group, specific aspects of the diagnostic strategy are still debated.

What is the role of occult AF and how do you rule it out?

The goal of cardiac monitoring is to identify patients with AF who need anticoagulation; the recommended length of recording is one of the most vigorously debated diagnostic challenges for CS. In particular, there is still controversy regarding the pathogenic role of short runs of AF.

The Event Monitor Belt for Recording Atrial Fibrillation After a Cerebral Ischemic Event (EMBRACE) study⁸ analyzed 572 patients (mean age in intervention group was 72.5 years, 54% were men, and median CHADS2 was 3), randomized to either 24-hour monitoring or 30-day recording, and found brief paroxysms of AF in 16.1% of those who had long monitoring as opposed to 3.2% of patients who had routine care. As a result, the investigators found that anticoagulant therapy was prescribed significantly more often in the patients who had the 30-day recording.

A new recommendation was published in the 2014 AHA/ASA guidelines stating that for patients who have experienced an acute ischemic stroke or TIA with no other apparent cause, prolonged rhythm monitoring (for about 30 days) to rule out AF is reasonable within 6 months of the index event (Class IIa; level of evidence C).⁹ Occult AF in the Asymptomatic Atrial Fibrillation and Stroke Evaluation in Pacemaker Patients and the Atrial Fibrillation Reduction Atrial Pacing Trial (ASSERT)¹⁰ was found in 10% of the 2,580 enrolled participants, who were monitored for 3 months. Of these patients with subclinical AF, mean age was 77 years, 56% were male, and mean CHADS2 was 2.2. Occult AF was independently associated with an increase by a factor of 2.5 in the risk of ischemic stroke or systemic embolism; this risk was independent of other risk factors for stroke and of the presence of clinical AF. Atrial tachycardia/fibrillation lasting longer than 5 minutes was newly found using implantable device in 30% of stroke patients during a follow-up of 1.1 ± 0.7 years with a median time to detection of 72 days.¹¹

The Cryptogenic Stroke and underlying Atrial Fibrillation (CRYSTAL AF) trial¹² randomized 441 patients to implantable cardiac monitoring (ICM) vs routine 24-hour recording and found that the rate of detection of AF with ICM was higher at 6 months (8.9% vs 1.4%; hazard ratio [HR] 6.4; 95% confidence interval [CI] 1.9–21.7; p < 0.001), 12 months (12.4 vs 2.0%; HR 7.3; 95% CI 2.6–20.8; p < 0.001), and 36 months (30.0% vs 3.0%; HR 8.8; 95% CI 3.5–22.2; p < 0.001). For the 168 patients randomized to ICM who had data available for the first year of recording, mean age was 61.3 years, 68% were men, and mean CHADS2 score was 2.9. The mean AF duration was 6 minutes or less in more than half the study sample and the majority of AF paroxysms were subclinical.

How do you rule out occult cardiac sources?

The role of thrombophilic states: Factor V Leiden mutation and antiphospholipid (aPL) antibody syndrome

Factor V Leiden mutation is mainly associated with venous thrombosis, not arterial, and anticoagulation is usually indicated when the mutation is homozygous. A study compared 44 young patients with CS with 282 healthy controls and found a higher prevalence of factor V Leiden mutation among stroke patients. The prevalence of all other 13 gene polymorphisms did not differ significantly.¹⁷ aPL antibody syndrome is more commonly associated with arterial infarction. The diagnosis requires a combination of clinical and laboratory criteria. As the Antiphospholipid Antibodies and Stroke Study (APASS) trial showed, the presence of aPL among patients with ischemic stroke does not predict an increased risk for subsequent vascular occlusive events over 2 years or a different response to aspirin or warfarin therapy.¹⁸

The usefulness of a screening for a thrombophilic state in patients with stroke or TIA is unknown (Class IIb; level of evidence C).

Expert opinion: Peter Rothwell, MD, PhD, FMedSci

The diagnostic approach for CS often varies from one center to the other and it is difficult to say what is cost-effective in general. One example is the monitoring strategies for AF: in our institution, we generally monitor patients for 5 days. Studies⁹ found that short runs of AF captured by long-term continuous monitoring with an insertable device are common. The significance of these short runs of AF remains uncertain and current excitement in the research community might be tempered if we think back to the misplaced emphasis that was placed on previous apparent causes of stroke, such as mitral valve prolapse. Prospective cohort studies are needed to determine if they are associated with an increased risk of stroke and if they are, then we need to establish whether that link is likely to be causal by performing randomized controlled trials of anticoagulation in these patients. Therefore, long-term monitoring should not yet become the gold standard for diagnosis, because it is difficult to recommend a specific treatment based on the available results. Using anticoagulation seems more defensive medicine than evidence-based, as there is controversy about whether this therapy is more efficacious than antiplatelet agents in these cases. Similarly, not all patients should get a TEE. Anatomical abnormalities such as left atrial enlargement can be successfully studied with a TTE and PFO and other shunts can be evaluated with a bubble study, so I would reserve the more invasive approach to selected cases. If, for example, a patient has a stroke following a long flight and the onset of stroke is during a Valsalva maneuver, then there is a strong likelihood of paradoxical embolism and one would look for a PFO. Moreover, ruling out complex aortic atheroma may not be useful for management because even if it is found, there is no good evidence to support anticoagulation vs antiplatelet therapy.¹⁹ That is also why specialized imaging such as cardiac MRI or multidetector CT is interesting for research purposes but not applicable in the clinical setting. In terms of screening for hypercoagulable states in case of CS, I have seen 3 different strategies (none of which has strong supportive evidence): (1) screen anybody irrespective of the age, but in this case you may find minor abnormalities of unclear significance, particularly in older patients; (2) screen patients who are younger than 45 years; (3) screen patients who are younger than 55 years, which is our approach.

Speaking of research, there are at least 3 topics that would merit further investigation. (1) The role of PFO in CS in the elderly (a previous study²⁰ showed a possible association and needs to be replicated). (2) The role of intracranial stenosis (IS) in CS, which may be underestimated. Many centers in Europe do not routinely obtain an intracranial MRA or CTA and limit the vascular imaging to carotid Doppler because they argue that the incidence of IS is low in the Caucasian population. We disagree and usually perform intracranial vascular imaging and found that many older patients have significant IS. (3) The link between migraine and CS. A potential shared mechanism may help to shed light on the pathogenesis of CS, possibly relating to intermittent abnormalities of the platelet function and endothelial activation with production of microemboli.^21,22

Preliminary survey results (April 20, 2016; 3 pm EST)

A total of 124 neurologists participated in the survey thus far, the majority of whom have a hospital-based (n = 43; 35%) or academic-based (n = 37; 30%) practice. Fifty-five percent of survey takers practice outside of the United States, with India and Italy being the most active countries (17% and 12% of responses, respectively).

Despite meta-analyses highlighting a great variability among studies that evaluated the role of TEE in patients with CS, the vast majority of readers (n = 94; 76%) favor this diagnostic technique in case of negative TTE. While it is possible that this result may vary depending on the clinical scenario in which CS is contextualized, this result highlights the need for randomized controlled trials to evaluate anticoagulation vs antiplatelet agents based on TEE findings.

As anticipated, there is no consensus regarding the length of monitoring to rule out AF: 25% (n = 31) of readers obtain a 5-day Holter, 29% (n = 36) obtain a 21–30 day Holter, and 22% (n = 27) utilize a long-term monitoring with implantable device. It is conceivable that these percentages will remain stable and the question unanswered until new evidence becomes available.