Transient Ischemic Attacks

FigureA 62-year-old man presents with a history of right-sided weakness, lasting probably less than five to seven minutes. He isn't sure, but his wife remembers that he had slurred, incomprehensible speech for at least a couple of minutes. They drove to the ED, and because he feels fine now, they wonder whether they should have come. He has a history of diabetes and of high blood pressure, both of which are reasonably well controlled. The emergency physician decides to get a head CT, which is normal. His bedside blood glucose is 102 mg/dl. You hear no carotid bruits, and refer him to his primary care doctor for follow-up. Three weeks later and two days after the couple return from a Caribbean cruise, he suffers a dense stroke. During his hospitalization, he is noted to be in atrial fibrillation. A lawsuit is filed against the emergency physician. One plaintiff's expert faults the EP for not obtaining an EKG and starting warfarin therapy. Another plaintiff expert faults him for discharging the patient at all, for not obtaining an EKG, and for not starting heparin therapy as a bridge to warfarin therapy. Both are willing to attest to a degree of medical certainty that this man would be functioning normally had he received competent emergency care. At deposition, the EP does not recognize the gentleman. He is unable to speak or walk, and cannot be deposed. The wife shoots daggers at the EP during the whole affair. In my past two articles, I have reviewed some difficulties in diagnosing transient ischemic attack (TIA) and in defining the entity. Considering that the incidence of TIA in the general population is approximately 68 per 100,000 population (Stroke 1998;29:2109), or between 200,000 and 500,000 TIAs annually in the United States (N Engl J Med 2002;347:1687), TIAs are something the EP deals with routinely. These figures have been amplified by a recent article demonstrating that there have been approximately 2,969,000 visits to emergency departments in the United States for TIA over a 10-year period ending 2001: three visits per 1000 or 0.3 percent of all ED visits. (Acad Emerg Med 2006;13[6]:666.) The traditional definition — the abrupt onset of a neurologic deficit lasting less than 24 hours and caused by a focal vascular lesion — is antiquated, as we have seen. Studies using magnetic resonance imaging show that 50 percent of patients meeting the traditional definition of TIA have actually had infarctions. (Cerebrovasc Dis 2005;19:362.) Of those with an index TIA, the risk of stroke within 48 hours approximates four percent to five percent. (Stroke 2003;34:e138 and 2005;36:720.) How Much Work-up in the ED? This question has a number of different answers, depending on the technology available and the timeliness of follow-up. In a recent review, 53 percent of TIA patients arrived between 8 a.m. and 4 p.m., an unusual excess of daytime visits. (Acad Emerg Med 2006;13[6]:666.) This figure is striking in that it may indicate that patients may not take symptoms seriously if they resolve quickly; after all, they didn't come in during off hours. EPs may not take the symptoms so seriously either. Once in the ED, head CT was obtained in 56 percent of patients and MRI in fewer than five percent of TIA patients. This is in spite of the fact that the American Heart Association recommends that head CT be performed on all these patients. (Stroke 1999;30:2502.) This also doesn't take into account that MRI with diffusion-weighted imaging is our best test to distinguish stroke from TIA. Of course, the lack of MRIs ordered is mitigated by the fact that this test is not available in most EDs. It is less understandable that electrocardiograms were only performed on 73 percent of TIA patients acutely. This is a low-tech test, available uniformly, and recommended by the American Heart Association (Stroke 1999;30:2502) and the National Stroke Association (Stroke 1999;30:1991) to be performed on all TIA patients. Atrial fibrillation is a major etiology for TIA, and treatment with heparin is recommended for new TIA patients due to atrial fibrillation. Back to Basics What else should be done universally? It doesn't make sense to perform expensive testing on inpatients or outpatients when there is no emergency treatment dependent on the test results. Antiplatelet agents are recommended by the AHA and the National Stroke Association. This may include aspirin or, if aspirin is contraindicated or has failed, ticlopidine or clopidogrel. Increasing the dose of aspirin has not been shown effective. The decision to use antiplatelet agents should be made in concert with the primary care physician or neurologist who will assume care for the patient after the emergency visit. In a recent 10-year review of TIA cases in the U.S., only 42 percent of TIA patients received any medications at all. No randomized, controlled study has ever shown the benefits of hospitalization for TIA. (Ann Emerg Med 2004;43:592.) This statement may not justify patient discharge, however. The major downside to not working up patients promptly in the hospital is the risk of missing occlusive carotid artery disease. The effectiveness of carotid endarterectomy is a function of the degree of stenosis and of the time delay that occurs between the TIA and surgery. (Lancet 2004;363(9413):915.) Carotid ultrasonography is noninvasive, with a sensitivity of 93% for high-grade stenosis and a specificity of 89% using cerebral angiography as the gold standard. (Mayo Clin Proc 1994;69:1027.) Echocardiography is noninvasive and may pick up left ventricular or atrial thrombus. Transesophageal echo may diagnose patent foramen ovale, aortic arch plaque, or atrial septal aneurysm, but is a more expensive and invasive test. I could find no literature that suggests that hospitalization is mandatory to accomplish ultrasound studies. Risk Stratification Can we risk-stratify which TIA patients with negative CT are at high risk of stroke? One review listed four groups possibly at high risk that might warrant immediate admission for further evaluation and treatment: ▪ “Aspirin failure” TIA. ▪ Possible cardioembolic stroke (i.e., atrial fibrillation). ▪ “Crescendo” transient ischemic attack: more than three ischemic events within 72 hours, with increased severity and frequency of symptoms. ▪ Patients with motor or speech deficits of more than 10 minutes, those 60 or older, or those with diabetes are at higher risk for early stroke. Clinical risk stratification models for TIA have been developed based on how many of these five findings were present. (Ann Emerg Med 2004;43:592.) The Bottom Line It appears there are approximately 10,000 to 25,000 patients who sustain strokes within 48 hours of an index TIA. With admission rates at only 54 percent for TIA, it is clear that EPs are discharging thousands of people from emergency departments who have really sustained strokes. Thousands more have sustained TIAs and will have a stroke within days of discharge from the emergency department, without a work-up meeting standards established by national groups. If there is a standard of care for TIA as defined by the American Heart Association or the National Stroke Association, then a lot of emergency physicians out there are not adhering to it. TIAs represent another entity in which standard of care may not be so standard. If the concept of acute ischemic cerebrovascular syndrome becomes accepted, then TIA will be seen within the spectrum of stroke, similarly to DVT and pulmonary embolus. If this is so, then expect that patients with TIA in the foreseeable future will be confined routinely in the hospital or observation bed pending rule-out for carotid artery disease or treatable cardiac etiology. The Bottom Line on Stroke Approximately 10,000 to 25,000 patients sustain strokes within 48 hours of an index TIA. ▪ Admission rates are at only 54 percent for TIA, meaning that EPs are discharging thousands of people from emergency departments who have sustained strokes. Thousands more have sustained TIAs and will have a stroke within days of discharge from the emergency department. ▪ Emergency physicians are not adhering to a standard of care for TIA as defined by the American Heart Association or the National Stroke Association.