Cryptogenic stroke

This recent Clinical Practice piece in NEJM is a difficult and cumbersome read. Here is my effort to distill it down to some key points.

The definition of cryptogenic stroke (CS) is unclear.

CS is generally defined as stroke which remains unexplained as to cause after adequate evaluation. However that concept is vague due to uncertainty about what constitutes an adequate evaluation. Moreover, some consider uncertain presentations with more than one plausible substrate identified as CS. The latter category is not considered CS for purposes of discussion in the paper. In an attempt at clarity the authors subdivide CS into 1) unknown cause after “standard evaluation” (CSSE) and 2) unknown cause after “advanced evaluation” (CSAE).

Standard evaluation is defined as: echo (TTE and/or TEE depending on clinical circumstances), in hospital telemetry or 24 hour Holter, stroke imaging (it doesn't specify whether this is CT and/or MRI), and noninvasive vascular imaging of the head and neck (CTA or MRA). Advanced evaluation consists of: invasive angio, transcranial doppler, vasculitis tests, outpatient monitoring up to 4 weeks and thrombophilia tests (arterial thrombophilia tests in unselected patients but venous thrombophilia tests only in those with suspected intracardiac shunt). A third category called “specialized evaluation” exists. It consists of: genetic tests, LP, brain biopsy, advanced cardiac testing and 1 to 3 year monitoring via an implanted monitor.

Causes of CS

A long list of rare causes is contained in the on line supplement to the paper. Here are the more important ones:

Occult atherosclerosis (nonstenosing but unstable plaques, intracranial and extracranial)

Thoracic origin stenosis

Nonatherosclerotic arteriopathies (eg dissection, vasculitis)

Hypercoagulable states (arterial in unselected cases, venous if paradoxical embolism is considered)

Low burden atrial fibrillation

Dilated cardiomyopathy

Paradoxical embolism

Likely causes by age

Realizing considerable overlap they can be categorized as follows:

18-30: Dissection is most common. Also consider thrombophilia and structural (congenital) heart disease.

31-60: Premature atherosclerosis, structural heart disease

Over 60: occult atrial fibrillation

Published stroke classification schemes may guide evaluation.

The CCS (Causative classification of strokes) (this is a useful web based tool) and ASCOD (atherosclerosis, small vessle disease, cardiac pathology, other causes, dissection) are emerging as more useful classifications than the traditional TOAST categories because they assign fewer cases as CS.

Recommendations for evaluation of CS

Recommended tests for the various levels of evaluation were listed in the first section above. Certain points warrant elaboration:

No particular order of testing was given within categories.

Small deep white matter infarcts, which might otherwise be categorized as CS, may not need advanced testing in patients over 50 or who have standard vascular risk factors.

Recommended vasculitis serologic tests consist of ESR, CRP, ANA, RF and ANCA.

Transcranial doppler monitoring for up to 60 minutes may detect covert microemboli, either of cardiac or arterial origin.

Invasive, as compared to noninvasive, angiography better evaluates medium and small arteries and may be useful in detecting vasculitis (which may be associated with negative serologic tests) and other unusual arterial diseases.

Special considerations in atrial fibrillation

These are detailed in the appendix available on line. Some of the recommendations differ from the secondary prevention guidelines.

The author recommends that when AF is detected during hospitalization anticoagulation with a NOAC be started by the time of discharge. This is a more aggressive approach than found in the guidelines and would raise some bleeding concerns in my mind. The guidelines, in contrast to the review article, state no preference between warfarin and NOACs other than to say that rivaroxaban has a IIa recommendation while warfarin, dabigatran and apixaban have a class I recommendation. Moreover, the guideline recommended start time is “within 14 days” in many strokes but delayed beyond that time in large infarcts or those with increased bleeding risk. Given the short hospital stay for most strokes these days the review author seems to favor anticoagulation significantly earlier than what the guidelines call for. Moreover, given the rapid onset of action of NOACs this approach would have the patient fully anticoagulated just a few days after stroke onset.

For CSSE the paper recommends discharge with ambulatory monitoring for 30 days, and on a NOAC if there are multiple AF risk factors but on antiplatelets if not.

For patients with negative 30 day monitoring and no AF risk factors the author would dismiss AF from further consideration and treat with antiplatelet therapy. If AF risk factors are present the author would recommend an implanted monitor with consideration for continuing the NOAC for a year followed by reassessment.

If low burden AF is found on 30 day monitoring the recommendation is to continue the NOAC but if no AF risk factors are present to consider stopping it after a year.

Further reading: here is a CS resource on the American Stroke Association website.