Anticoagulation for secondary prevention of stroke

As an internal medicine resident, I often let my neurology colleagues make decisions about patients’ anticoagulation after a stroke. However, in clinic, I recently had a 70-year old man who had immigrated from Taiwan who had a history of vertebrobasiliar infarcts, Parkinsonism, and coronary artery disease who presented to establish care. He was on clopidogrel. His wife reported that his gait was poor and that he fell at least 5-6 times a week, sometimes hitting his head on the stairs or a table. His medication compliance was poor.

This led me to wonder: what agent should be given for anticoagulation for secondary prevention of stroke?

Options include:

  • aspirin
  • clopidogrel
  • dipyridamole
  • DOACs
  • ticlodipine (don’t use ticlodipine)
  • cilostazol (only studied in Asian populations, more expensive in the US)

Should you choose aspirin or clopidogrel? 

It depends greatly on the kind of stroke (was there major stenosis? Lacunar infarct? A minor ischemic stroke or TIA?) as well as the patient (are they elderly? Have diabetes? Atrial fibrillation?). The combination of aspirin + clopidogrel can be used for the first 90 days for ischemic stroke/TIA, but beyond that, this combination is not superior to aspirin or clopidogrel monotherapy, and causes higher rates of bleeding, as you might expect. CAPRIE showed that clopidogrel was more efficacious than aspirin; however, this was only significant for patients with PAD.

If you choose aspirin, 81 mg or 325 mg? 

This review nicely shows that doses of aspirin used in trials vary widely (from 30 mg to 1300 mg) and that for most patients, using 75-81 mg aspirin is probably the sweet spot to provide adequate protection while minimizing risk of bleeding.

I’ve heard about an aspirin-dipyridamole combination…? 

ESPS-1 and ESPRIT showed that the combination of aspirin + dipyridamole resulted in better outcomes than aspirin alone for large vessel disease. The recent PROFESS trial showed this combo is comparable to clopidogrel monotherapy. The caveat is that patients often have side effects from dipyridamole like nausea and fatigue and are not able to tolerate it.

Are there special considerations for patients with atrial fibrillation? 

There is now evidence suggesting warfarin, apixaban, and dabigatran are also ok in patients with AF who previously had strokes. (Class I, grade B recommendation) . Two birds with one stone!

I ended up switching my patient above to 81 mg aspirin. With his CAD, I felt that it was important for him to be on aspirin, and he didn’t have stents or other disease that would necessitate clopidogrel. Furthermore, with his falls and poor medication compliance, I didn’t feel that an irreversible agent or adding on dipyridamole would provide much benefit.

Is is safe to give tPA/thrombolytics to someone who just underwent cardiac catheterization?

At my institution, we recently discussed a case of a patient with acute stroke in the setting of cardiac catheterization (right femoral access). The decision to push tPA was a tricky one, because he had just been cathed. When is it ok? When is it NOT ok?

Here are relative and absolute contraindications (respectively) for giving tPA:

 

As you can see, arterial puncture at a noncompressible site is a relative contraindication for tPA. So depending on where the puncture site was…the typical site for an arterial stick in a catheterization is between the bifurcation of the SFA and PFA and the branching point of the inferior epigastric artery. Typical site=compressible, “high stick”=noncompressible. If there is a “high stick,” tPA will confer a much higher risk of RP bleed and hemorrhage, and should be avoided.

Screen Shot 2018-02-06 at 11.54.23 AM

Should I be worried that this patient has a stroke mimic, and is it still safe to push tPA/thrombolytics?

When you’re called about a patient with new dysarthria, weakness, or altered mental status, stroke is one of the first and most dangerous things that probably comes to mind. Stroke, whether ischemic or hemorrhagic, can present in a variety of different ways depending on the particular insult–anything from tingling in a few fingers to full-on obtundation.

There are quite a few things that mimic stroke (see below), but you should always assume that an acute neurological deficit might be stroke unless there is a very convincing alternative explanation. Stroke is a clinical diagnosis, which means that you do NOT absolutely need brain imaging to prove it. In this “brain attack” study from Scotland, published in 2006, features associated with true strokes were: patient was well within the last week, an exact time of onset can be stated, obvious focal deficit is present, NIHSS score >10 (although some mimics did have high NIHSS scores as well). On the other hand, stroke was less likely in patients with a history of cognitive disturbance, loss of consciousness at onset of episode, witnessed seizure, lack of lateralizing deficits, and who were still able to walk. This study was observational and did have significant limitations–for example, not all patients got imaging and there was no gold standard defined for diagnosis of stroke–but these features square with what I’ve been taught.

Stroke mimics: 

  • TIA (although many high-risk TIAs are later found to be strokes)
  • Stroke recrusdescence (stroke-like symptoms in a patient with a history of stroke triggered by an acute cause)
  • Dementia or delirium (including toxic-metabolic encephalopathy)
  • Seizure
  • Space-occupying lesion (infection, tumor)
  • Syncope
  • Vestibular dysfunction
  • Migraine
  • Spinal cord lesion
  • Peripheral mononeuropathy

What if you suspect that your patient might actually have a stroke mimic, but you can’t get imaging in time (<3 hours) and they have concerning neurological symptoms and a high NIHSS? Should you withhold tPA? This meta-analysis from Greece suggests that patients who have a stroke mimic also have a lower risk of bleeding from tPA, so if there is any concern about an ischemic stroke, this study recommends giving tPA.

What happens during table tilt testing?

Tilt table testing is part of the work up for vasovagal syncope. It is also indicated in:

  • recurrent episodes of syncope when a cardiac cause has been ruled out
  • unexplained syncopal episode in a high-risk setting
  • to examine reflex syncope (and distinguish between orthostatic and reflex syncope)

I saw a case of a 26 year old who passed out after dinner with friends. According to her friends, she was shaking a little but came to spontaneously after a few minutes. Because her EKG and Holter and EEG were normal, she got a tilt table test. She actually started reporting lightheadedness after getting SL nitro, and she had to get fluids and put in the supine position. Her HR during the test was all over the place, 90s-130s, and changed really quickly, like 90 one second and then 102 the next, then 96, then 113. Dr. Sood said that people’s HR do change, but it’s usually more gradual and not that extreme, so this qualifies as sinus arrhythmia—she has primary sinus tachycardia. He postulated that her syncope was due to tachycardia itself, so put her on a low-dose beta blocker.

Tilt table testing has limited sensitivity and specificity. Some report that an individual can have a negative result one day, and then a positive result the next. But it is useful for drawing the following conclusions:

  • If reflex hypotension or bradycardia + syncope are induced, the test is diagnostic for reflex syncope.
  • If there is a progressive, slow decrease in systolic blood pressure (orthostatic response) with or without symptoms, syncope due to orthostatic hypotension is diagnosed.
  • When syncope is induced but there is no hypotension or bradycardia, this suggests psychogenic pseudosyncope.

What do you assess for in brain death?

Doing exams for brain death is a grim task, and it’s important to be sure about your findings. If I could impart one tip, it would be: push harder than you need to, be louder than you should, and don’t do the exam half-heartedly. It will make you more confident about reporting on the patient’s status.

There is a very nice outline of the brain death exam with guidelines and numbers on UpToDate. Here are some clinical pearls I have learned along the way:

  • First off, brain death exams should be done serially, beginning at seventy-two hours
  • It’s called noxious stimuli for a reason. One attending I know has a technique she calls “the nipple twister.” Watch the pt’s movement, facial expression, BP, and heart rate to evaluate their response to pain
  • use a real flashlight to evaluate pupils
  • use a saline jet to evaluate the corneal reflex
  • really snap the head back and forth when doing a test for oculomotor reflexes. It’s different than turning the head to and fro for sound localization
  • reflexes aren’t the best indicator of higher-level activity because they can still happen at the spinal cord level
  • the family can always be in the room if they want, but warn them that the exam may distress them. Many times, the families are still in shock while we’re evaluating their loved ones for brain death and it’s important to be respectful of their own pain.