Is this a vascular event?
Dr J.T. Butler, Dr J. Carr
Case 1: A 58-year-old hypertensive man suddenly develops left-sided weakness while watching television. On examination flaccid left hemiplegia, left hemisensory loss, hemianopia and mild drowsiness are evident. You think that he has had stroke due to thromboembolism of the right middle cerebral artery, but wonder how certain you can be of your diagnosis.
Case 2: A 50 year old, previously well man reports that on three occasions numbness and paraesthesiae of the left side of his body and head occurred without associated symptomatology. On 1 occasion it started in the left foot and spread slowly up to the left side over 15 minutes. The examination is normal. You wonder whether he has had transient ischaemic attacks, seizures or even multiple sclerosis.
Case 3: A 38-year-old hypertensive, diabetic patient acutely develops weakness of his right arm and mild clumsiness of the left hand, numbness of the right face, arm and leg, dysarthria, vertigo and mild occipital headache at midday. On examination his BP is 180/110, he has very poor tongue movement, moderate weakness of his right arm and impairment of light touch and pinprick of the right face, arm and leg. By the following morning all his signs have resolved and his blood pressure has diminished to 165/100 on his usual diuretic and ACE inhibitor. He still feels that his right hand is slightly clumsy. You consider that he has had a brainstem TIA or ischaemic stroke and request a CT prior to administering aspirin.
A cerebrovascular event traditionally has been used to describe the pathophysiologic entities of reversible ischaemia, infarction or haemorrhage of the central nervous system. Mechanisms of cerebral ischaemia include arterial embolism, thrombosis, thromboembolism, small vessel disease and haemodynamic disturbances that may involve a myriad of different blood vessels of different sizes. These processes may be a consequence of hypertension, diabetes, atherosclerosis, valvular heart disease, cardiac dysrhythmias or cardiac wall motion dysfunction. Uncommon causes include vasculitides, coagulation disorders, atrial myxoma and paradoxical embolism. Rarely venous infarction of the brain may occur as a consequence of venous or sinus thrombosis. Haemorrhage may occur into the subarachnoid or subdural space or within the brain parenchyma and may be due to hypertension, aneurysms, vascular malformations, trauma, coagulopathies or occur without clear cause.
Traditional views of the diagnosis of cerebrovascular events
To the clinician this bewildering complexity may seem daunting. Standard texts have brought solace to the anxious physician by painting the clinical presentation of cerebrovascular disease as having:
Opinions about the definition or diagnosis of stroke abound. One expert defines stroke as "sudden, non-convulsive, focal neurologic deficit", while another states that "in most cases, the abrupt, dramatic onset of focal neurologic symptoms stamps the process as a stroke, particularly when the symptoms correspond to a specific vascular territory" (1,2). Yet another expert says that "the clinical picture of stroke, either due to cerebral infarction or indeed to primary intracerebral haemorrhage, is characteristic and the diagnosis depends not so much on where the lesion is thought to be but on the sudden or subacute onset in a previously well patient with no history of head injury" (3).
A TIA is "an acute loss of focal cerebral or monocular function with symptoms lasting less than 24 hours". The diagnosis is usually made by history alone, based on symptoms that are identical to a stroke but only shorter in duration. As discussed below, the reliance on the history is the cause of the greater error rate of diagnosing TIA’s than strokes. Furthermore, it is the brevity of the symptoms which causes TIA’s to be mimicked by seizures, migraine, multiple sclerosis, hypoglycaemia and shunting in vascular malformations. The experienced clinician uses various strategies to improve diagnostic accuracy:
Is the distinction of TIA from stroke valid?
The distinction of transient ischaemic attack (TIA) from stroke using a cut-off time of 24 hours for the resolution of neurologic deficit had as its explicit purpose the separation of reversible cerebral ischaemia from irreversible neurologic dysfunction (4). However, the diagnosis of TIA, while suggestive of transient or reversible ischaemia, does not exclude infarction on computed tomography (CT) or magnetic resonance imaging (MRI) (5). Likewise, features of previous infarction are commonly seen on CT or MRI or patients who are not aware of having had a stroke (subclinical infarction).
Rarely a TIA may represent a haemorrhage (a small bleed or ischaemic shunting in a patient who harbours a vascular malformation may present as transient neurologic dysfunction). As ischaemic neurologic events are very much more common than vascular malformations in patients who present with TIA’s, it is likely that the clinician will be correct more often than not when concluding that there is an ischaemic basis for a clinical diagnosis of TIA.
Precision of the diagnosis of TIA
A variety of conditions may mimic TIAs, including seizures, migraine, vascular malformations, haemorrhages and hypoglycaemia. As the diagnosis is most commonly made by history alone, potential sources of error include patient recall, the patient's interpretation of symptoms, language difficulties, patient education and intelligence, meticulousness with which the history is obtained and the interpretation of the symptoms by the physician. Errors in diagnosis have practical implications as aspirin, ticlopidine, dipyridamole and carotid endarterectomy have all been shown to be of benefit in selected patients for preventing further TIAs or stroke.
When experienced physicians are asked to decide on whether a "previous ischaemic episode" has occurred in patients they have seen in consultation, only fair agreement ensues (6,7). In the absence of a reliable "gold standard", the accuracy of a history of TIA for predicting the existence of a state of reversible cerebral ischaemia is not known.
In conclusion, the clinician has to accept some uncertainty when making the diagnosis of a TIA.
Carotid versus vertebrobasilar TIA?
Distinguishing carotid from vertebro-basilar artery territory TIA's is of considerable clinical importance, as patients with the latter would not be expected to benefit from carotid endarterectomy while carotid territory TIA's may require work-up to exclude the possibility of significant carotid stenosis, requiring endarterectomy.
The carotid arteries supply much of the cerebral hemispheres but not the cerebellum, brainstem or primary visual areas in the occipital cortex. The symptom of an isolated hemianopia strongly implicates the primary visual area of the contralateral occipital lobe, whereas monocular visual loss (amaurosis fugax) lies within the carotid artery territory. Brainstem and cerebellar symptoms include diplopia, vertigo, tinnitus, slurring of speech, difficulty with swallowing, imbalance and the illusion that objects are moving (which is the subjective experience of nystagmus, also known as oscillopsia). As the brainstem contains ascending and descending tracts that connect the cerebral hemispheres with the spinal cord, weakness and sensory symptoms (numbness and paraesthesiae) may be associated with the aforementioned brainstem and cerebellar symptoms in vertebrobasilar TIAs. The greater the number of these symptoms occurring simultaneously in any patient, the greater the confidence of deciding that the event is in the area supplied by the vertebrobasilar system.
Experts presented with the same clinical information about patients who have had TIA’s have poor agreement about the territory involved (6). This is strong evidence that clinicians' diagnoses of the involved territory are frequently incorrect. Caution is therefore necessary when deciding whether a patient who has experienced a TIA should undergo evaluation for carotid endarterectomy or not.
Accuracy of the diagnosis of stroke
When emergency unit clinicians use the definition of stroke as a history of "persistent focal neurological deficit of acute onset during the prior week (but without head trauma)", accuracy is high (sensitivity = 86% and specificity = 99% in one study). It is likely that general practitioners using this definition would obtain a similar accuracy, perhaps with some loss of sensitivity (as a result of seeing patients earlier and with milder disease) and some diminution of specificity (because of less experience with stroke). In summary, a few strokes will be missed and very few inadvertently labelled as other entities. This diagnostic error rate is considered low when compared to other everyday clinical diagnoses and the consequences are not enormous.
When the above definition is used for patients admitted to hospital, specificity diminishes because of the referral bias of complicated cases (abscesses and tumours that bleed may mimic a stroke, resulting in "false positives" (6).
Mimics of stroke
The following may present with a sudden onset of neurological deficit, and therefore mimic stroke (8):
Precision of stroke classification
Agreement among experts is poor when attempting to classify stroke subtypes (e.g. cardiac embolism, large vessel atherosclerosis, lacunar stroke, haemorrhage, and subarachnoid haemorrhage) (6). This strongly suggests that clinicians are not accurate in classifying stroke subtypes. When using clinical and neuroimaging findings, it is frequently impossible to draw a distinction between ischaemic stroke subtypes (such as embolic stroke arising from a cardiac source or occlusion of a large vessel associated with focal atheroma).
When using clinical information, experts are modestly adept at distinguishing ischaemic from haemorrhagic stroke: e.g. of patients with haemorrhagic stroke, 48-76% are correctly identified while 68-96% with ischaemic stroke are correctly identified (6). This too has implications for therapy, as neuroimaging would be required prior to the use of antiplatelet or anticoagulant therapy.
CT scans reliably identify acute intracerebral haemorrhages and should therefore be requested when diagnostic doubt exists about stroke subtype, particularly where decisions about the use of antiplatelet or anticoagulant therapy are required. CT scans may be normal within the first few hours following ischaemic stroke. Occasional patients who harbour vascular malformations, haemorrhagic tumours and abscesses will readily be identified. In someone who clinically is considered to have suffered an ischaemic stroke, the greatest virtue of CT scanning is probably it's ability to exclude or rule out haemorrhage and other mimickers.
Resolution of clinical scenario
Case 1: You consider it extremely likely that he has had a stroke but because you wish to consider him for aspirin therapy, you refer him for a CT scan to exclude a haemorrhage. The CT demonstrates early ischaemic change in the right middle cerebral artery, prompting you to gradually reduce his blood pressure, to give aspirin 150 mg daily, to apply stockings for deep vein thrombosis (DVT) prophylaxis, to ask for physiotherapy to preview pneumonia and DVT and to use a nasogastric tube.
Case 2: You consider that sensory seizures rarely last 30 minutes and that the events were not perfectly stereotyped. An EEG is normal. You therefore consider the probability of seizures to be small. The probability of multiple sclerosis or a vascular malformation appears low because of their relative infrequency. The blood glucose is 11.7, enhancing your suspicion that he has had TIA’s (diabetes a risk factor for focal cerebral ischaemia). You refer him to a physician, who requests a CT because of his uncertainty. This demonstrates a small infarct of the right lateral thalamic region, almost certainly causally related to the symptomotology.
Case 3: To your surprise the CT scan shows a pontine haemorrhage (see fig. 1), prompting you to avoid aspirin and to treat his blood pressure.
In a nutshell
medicine. Eleventh Edition. McGraw-Hill, Inc, 1987:1932
I. Which of the following symptoms increase the probability of a vertebrobasilar (rather than a carotid) territory stroke
Answer - 4
II. Which of the following are true
Answer - 3
Which of the following is incorrect
5. When patients with suspected stroke are subjected to neuroimaging because of diagnostic uncertainty, CT scan should suffice in most instances.
Answer - 3