Stroke vs Stroke Mimics: Diagnosis at the Bedside

Abstract & Commentary

By Dana Leifer, MD, Associate Professor, Neurology, Weill Medical College of Cornell University. Dr. Leifer reports no financial relationship related to this field of study.

Synopsis: This prospective study demonstrates that clinical features such as focal deficits, a clear time of onset, and absence of non-neurological signs distinguish a stroke diagnosis from other diagnoses at the bedside.

Source: Hand PJ, et al. Distinguishing Between Stroke and Mimic at the Bedside: The Brain Attack Study. Stroke. 2006;37;769-775.

The availability of a growing number of therapeutic options for acute stroke patients makes rapid and reliable diagnosis of stroke at the bedside more important than ever. Modern technology, such as CT and MRI, can often diagnose stroke and rule out other conditions, but to avoid wasting time and resources, accurate diagnosis must be made efficiently by emergency medical personnel in the field and by physicians and nurses in the emergency room.

Hand and colleagues prospectively studied 336 consecutive patients with suspected stroke. Patients were identified by emergency room personnel as soon as possible after arrival and by review of admission registers from the emergency room, stroke unit, and neurology ward. Clinical evaluations were performed by neurology or internal medicine residents, and final diagnoses were determined by the consensus opinion of a panel of experts.

With 350 acute events in 336 patients, the final diagnosis was stroke in 241 cases and stroke mimic in 109, which included 44 episodes that were considered possible stroke or transient ischemic attack. Sixty-two of the stroke mimics were seen within 6 hours. This is an important group because it includes most patients eligible for intravenous thrombolysis or intra-arterial interventions. In this group, seizures accounted for 29% of the diagnoses, syncope in 14.5%, sepsis in 9.7%, toxic/metabolic changes in 9.7%, acute mononeuropathy in 6.5%, space-occupying lesions in 4.8%, acute confusion in 4.8%, vestibular dysfunction in 4.8%, dementia in 3.2%, and migraine in 3.2%. In patients presenting after 6 hours, seizures and syncope were less common and accounted for only 10.6% and 2.1% respectively, but sepsis and space-occupying lesions were more frequent, accounting for 17.0% and 14.9% respectively.

Univariate analysis demonstrated that patients with an uncertain time of onset, seizure at onset, loss of consciousness, non-neurologic symptoms, prior cognitive impairment, no lateralizing symptoms or signs, or signs not consistent with symptoms, were less likely to have a stroke. In contrast, an exact time of onset and any focal neurologic sign or symptom (speech difficulty, visual loss, focal weakness or numbness, upper limb ataxia, extensor plantar) predicted a diagnosis of stroke, as did presence of coronary or peripheral vascular disease and hypertension (SBP > 150, DBP > 90). More severe deficits as measured by the NIH stroke scale (NIHSS) were more likely to be associated with a stroke, as were patients whose syndrome could be classified as a total or partial anterior circulation stroke by the Oxfordshire classification system. Vertigo and leg ataxia were not significant predictors because they occurred frequently in vestibular dysfunction.

Multivariate analysis identified presence of non-neurologic abnormalities and prior cognitive impairment as factors independently predicting that a patient did not have a stroke. Exact time of onset, definite history of focal neurologic symptoms, any abnormal vascular findings (SBP > 150, atrial fibrillation, valvular disease, or absent peripheral pulses), any lateralizing signs, and definite classification by the Oxfordshire system all predicted that a patient had a stroke. The multivariate analysis also confirmed that chance of stroke increased as the NIHSS score increased. The most powerful predictors were definite history of focal neurologic symptoms and NIHSS greater than 10.


These results are important because they demonstrate that a few key features make the diagnosis of stroke likely. The results of the study suggest that initial evaluation of potential stroke patients should determine if there is an exact time of symptom onset, any definite history of focal neurologic symptoms, and any lateralizing signs. The key symptoms and signs are straightforward—speech difficulty, visual loss, focal weakness or numbness, arm ataxia. If such symptoms or signs are identified, this study suggests that it is appropriate to activate a rapid protocol for more thorough evaluation by imaging and more detailed clinical evaluation focused on reaching a definite diagnosis and starting treatment.

On the other hand, if focal signs and symptoms are absent or non-neurologic problems are present, diagnoses other than stroke should be considered, and these include potentially serious non-neurologic conditions such as sepsis, syncope, and toxic-metabolic disorders. Although these conclusions may seem obvious to neurologists who have experience with stroke, the study makes a significant contribution because it provides guidelines for non-specialists who may see a stroke patient first, and must recognize that a patient may be having a stroke.